Use Of Ambient Light Sensors Smart Phones

Published Date: 02 Nov 2017

VELAYUTHAM CHIDAMBARAM

Abstract: In recent years Ambient Light sensors have become an integral part of smart phones in order to reduce the power consumption of the displays and enhance the user readability. This literature survey discusses about the different ALS topologies, their features and their implementation in various models. The ALS algorithms used in few of the most popular Ambient Light Sensors and their operation has been analyzed and few alternatives to Ambient Light Sensors have also been presented.

Keywords: Ambient Light Sensors, LED Drivers, Intelligent Sensors, Content Based Adaptive Control, Integrating ADC Convertor.

I. Introduction

In the current scenario a Smartphone has become an inevitable part of our lives. When people buy a Smartphone their major concern apart from the cost and features is the power consumption of the phone. Over the years the growth in the features of a Smartphone has seen a mighty growth but the power consumption in smart phones has not developed to such an extent to match these features thereby leading to reduced battery life . A recent survey shows that liquid crystal display and its associated backlighting are among the more and frequently power hungry components in a Smartphone [3]. Therefore for Smartphone manufacturers reducing the power consumption of these components in order to provide the customers with increased battery life has become one of their critical design considerations. The most prominent component that has been used by all Smartphone Providers has been the Ambient Light Sensors (ALS), these sensors sense the ambient light and accordingly control the operation of the backlight LED based on those inputs. The Ambient Light sensor also makes sure that the display is illuminated in such a way that it is readable to the user. When the Smartphone is used outdoors on a sunny day, the ALS senses the ambient light and accordingly transmits a signal which would cause the LED Backlight to increase its brightness, whereas on a dark environment the data measured by the ALS would cause the LED backlight to be dimmed [11]. The basic operation of a ALS is described by the block diagram in Fig 1. [1]

LED Driver / Microcontroller

ALS Algorithm

ALS

Adjust the Brightness

Figure 1: Block Diagram Representing the basic operation of Ambient Light Sensor

II. Spectral Sensitivity Of ALS vs. Human Eye

The main criteria for an ALS should be its photopic characteristics i.e. its spectral response should be similar to human eye. Normal human eye has a spectral response ranging from 400 to 700nm, its peak intensity is at 560nm but standard silicon optical sensor detect a wider spectral response ranging from 350 - 1100 nm with its peak intensity centered around 880nm, thereby these sensors can sense IR radiations (peak intensity - 880nm) which are not detected by human eye and therefore would perceive much brightness than what the eye actually sees [2][14].

Ambient light sensors also use Silicon but have different chip structures that would enable the IR radiation to be compressed and have its peak intensity around 550nm so that it can perceive the same brightness as perceived by the human eye (Figure 2). The ability of the Ambient light sensor to be photopic i.e. have the same spectral response as that of the human eye is very essential for its operation as the control signals are generated are based on these inputs [14].

Figure 2. Spectral sensitivity of a standard silicon detector and an ambient-light sensor [11].

The photopic features of ALS can be improved by many other approaches such as [13]

1.The IR filter compounds are placed into the epoxy in order to reject the IR Light so that ALS would mimic a human eye.

2.Two Photodiodes can be used one to sense everything and the other to receive the IR radiations and subtract it from the first photodiode .

In order to improve the display quality of the Smartphone while allowing the ALS to operate efficiently certain manufacturers have come up with RGB ALS sensors that sense the color components of the ambient light and this data after being processed by the microcontroller allows the backlight to control the respective color components such that both power reduction and superior image quality is obtained [12].

III. Different ALS Topologies

Different types of optical sensors are available for the designers to choose from, they are 1.Photoelectric cells 2.Photodioded 3.Phototransistors 4.Photo ICs. All these optical sensors have certain tradeoffs which is listed out in the Table 1.Of all these optical sensors PhotoICs have many pros and overcome many of the disadvantages of the other optical sensors. Photo ICs have an integrated photodiode with additional features such as amplification, ability to shutdown , low dispersion ( mimics the human eye) and complex logic control. These photoICs are available in both Analog and Digital types.[3]

Optical Sensor

Advantages

Disadvantages

CdSPhotoelectric cell

Photopic

Contain Cadmium, a prohibited RoHS

Photodiode

1. Very low dispersion between individual units

2. Photopic

Produces a very small output signal

Phototransistor

Large Output Signal

1. High dispersion between individual units 2. Poor temperature characteristics

Photo IC

1. Photopic

2. Low Dispersion

3. High Output signal

4. Built In Shutdown feature

5. Reduced Circuitry

Table 1 : Comparison Of Different ALS Topologies [1][3]

Analog ICs :

Analog ALS produce an output current/voltage that is proportional to the ambient light sensed by the photodiode integrated within the IC. The voltage/current sensed drives the ADC on the Microcontroller which sends commands to the LED driver to control the backlighting operation. In few cases the LED driver IC is equipped with autonomous control thereby eliminating the need of a microcontroller.

Figure 3: Analog IC [3].

Digital ALS ICs:

The photodiode output signal is amplified, the integrated ADC (usually a 16-bit ADC is used) converts the optical sensor’s output to I2C signal which is communicated to the microcontroller through I2C interface. The microcontroller performs the control operation.

Figure 4: Digital IC [3].

Digital ALS have a simplified circuitry compared to their analog counterparts which makes it easier for the designers to integrate multiple sensors on a single two wire bus (I2C) in a digital ALS using minimum connections, whereas in Analog ALS each new ALS would require two wire connections. Compared to Analog ALS Digital ALS has a better noise rejection ratio. But the power consumption in both the Analog and Digital ICs are same. But due to Digital ALS’s many advantages it is the most preferred ALS topology in the current scenario

Sensor Type

Sleep Mode

Active Mode

Digital ALS

1µA

190 µA

Analog ALS

0.4 µA

97 µA

Table 2:Power Consumption in Analog ALS and Digital ALS in different modes [12].

The Table 2 draws a power consumption between the Digital and Analog ALS. Even though the Analog ALS consumes less power in the long run the power consumption of both the ALS are the same. These ALS can operate in the range of .05 lux up to 16K lux [12].

IV. Novel Methodologies Of Implementing Ambient Light Sensors

Nowadays there has been a widespread research going on in the Area of Ambient Light Sensors, many novel methods of implementing these sensors have been proposed and successfully integrated into the existing models. The Most effective and novel implementations will be discussed in the following text.

1.On -Chip Ambient Light Sensor in an LED Driver with constant Off Time.

Usually Ambient Light Sensors are placed outside the LED Driver (Control the LED Backlight Current), this approach consumes more circuit area on the PCB and greater power consumption because the supply current for ALS operation increases with increased ambient light brightness and the sensed signal may be contaminated. These drawbacks are overcome by integrating the ALS within a LED Driver [4].

In order to achieve minimum area on the PCB the ALS is placed in series with a MOSFET as it produces a better signal by utilizing low area. The ALS senses the ambient light and produces a equivalent reference voltage which is compared with the LED Backlight voltage. These are given as inputs to an Error Amplifier, when the output of the Error amplifier is exceeding the thresholds then it means the Backlight current is either dark/too bright compared to the ambient lighting. Therefore the LED driver sends the corresponding signal to either increase or decrease the backlight LED current. The below block diagram (Figure 5) gives a pictorial description of the ALS.

Figure 5: Basic Operation of an on-chip ALS in LED Driver [4]

In this approach since the ALS is placed within the LED Driver Circuit it is prone to various photo radiations from the other transistors in the circuit and noise, to eliminate these the ALS consists is provided with a metallic covering. The major problem in several modern ALS are the rejection of the dark current control. Dark current control majorly affects the operation of an ALS. The main source of dark current generation is dc voltage, hence here the dark current generated is duplicated by using a dummy photodiode and subtracted from the original ALS thereby enabling its smooth and efficient operation. In addition to these features this model is also equipped with the feature of setting a constant off time for the automatic ambient dimming control (similar to an interrupt). This off time is programmable and the operation of the loop controlling the backlight current is in sleep mode thereby consuming less power [5].

Experimental Results :

In this experiment the on chip ALS Led Driver was exposed to ambient light as low as 30 and as high as 750 lux. The average power consumption was as low as 100 mW in normal operation. Much low values of power were obtained by setting a constant off time to the LED Driver. The below Figure 6 depicts simulated waveforms between the output voltage sensed by the ALS proportional to the ambient light and the corresponding Led Backlight current [4].

Figure 6: Simulated waveforms of the ambient light lux, ALS voltage and LED Backlight current

2.Use Of Ambient Light Sensors to prevent Luminous Decay This approach aims at reducing the Luminous Decay (results due to aging of the LEDs) of the Led Backlight in addition to performing the basic operation of sending signals to the microcontroller to control the backlight in accordance with the ambient lighting conditions[6].

Architecture: The architecture of the proposed system is shown in the block diagram shown in Figure 7. There is a LED light source embedded with sensors to detect both the ambient light and the LED brightness. The ambient light signal sensed by the ALS is sent to the microcontroller which sends appropriate signals to the LED Driver and at the same time checks if there is any luminous decay within the LED tube. If there is any luminous decay then the brightness is accordingly adjusted (increased/decreased) without affecting user performance. Although this approach consumes slightly more power due to another ALS and additional circuitry, on the long run it is a better trade off as the longevity of the LED Backlight is increased. The block diagram below (Figure 7) represents the basic operation.

Figure 7: Basic Block Diagram representing the major components and basic operation [6]

The ALS that is used to detect the ambient light is placed external to the LED light source and the ALS used to detect the brightness of the LED tube is placed inside the LED Backlight tube consisting of an array of LED's that act as the backlight. There is a black mask between these two sensors to avoid any interference of the signals

Hardware: The main components that constitute the hardware of this model are 1.Microcontroller 2.ALS 3.ADC 4.LED Driver. The signals sensed by the ALS are converted to analog signals by the ADC and transformed to the microcontroller. The microcontroller does calculations based on these signals and transfers them to different duty cycles of PWM . Based on these PWM signals the LED driver controls the backlight. For instance when there is a luminous decay, the microcontroller increases the PWM duty cycles thereby causing the LED Driver to increase the average current output resulting in increased brightness [6].

Software: In this model the algorithm carried out is so detailed that the PWM control pulse send by the microcontroller based on the calculations to change the backlight settings is more accurate (3 - 100 steps increase ) compared to the normal 10 step brightness signal. The flow chart(Figure 8) below lists in detail the algorithm adopted

Figure 8: Flow Chart depicting the Algorithm [6].

The Brightness of each LED within the LED array is sensed by the microcontroller and compared with the other the other LED's. If the brightness of other LED's is greater than the first one it decreases the brightness of the first LED but increases the brightness of the other LED's to produce sufficient backlight for user readability . Suppose the LED's within the LED tube have reached maximum brightness then the brightness of all the LED's are reduced by 1%.

This approach had enabled the power consumption to be lowered by around 40 % and at the same time has increased the longevity of the LED tubes by 20% which is substantial improvement [6].

3.Ambient Light Sensors used in conjunction with a Proximity Sensor

Nowadays in most of the Smart Phones proximity sensors have become an integral part to reduce power consumption and enhance user experience. A Proximity sensor illuminates the surrounding environment by the IR light emitted from a photodiode and senses back the reflected signal. This signal is sent to the Microcontroller which determines if the bombarded object is human face or not and sends the signal to the LED Driver to darken the LED backlight when the Smartphone is brought near the human face (especially when a person is attending a call) [13].

Given the operation of a Proximity Sensor, the method of operation is quite similar to an ALS as both sense the signals using a photodiode and transfer it to the microcontroller after ADC conversion. So there is bound to be some interferences between the two signals sensed by these sensors. Therefore during the operation of an ALS, once the ALS senses a signal it is converted to a digital signal and sent to the microcontroller for control operations, the signal does not pass through a demodulator (important component in the operation of proximity sensor).

Similarly during the operation of a proximity sensor, apart from returning the reflected signal, the photodiode also generates large magnitude of constant current due to the ambient light, thus for effective operation the constant LED current generated due to the ambient light is actually detected and cancelled out before the demodulation phase. The below Figure 9 shows a circuit containing an ALS along with a Proximity Sensor [7].

Figure 9: An architecture comprising both Ambient Light Sensor and Proximity Sensor [13]

In this circuit the Photodiode array performs the ALS operation, the IR photodiode array performs the proximity sensor operation, and the reflected signal received by the IR photodiode array is subtracted from the Photodiode array to eliminate the constant current component generated from the ambient Light. These signals are then converted to digital signal by an ADC and transferred through the I2C interface or by means of interrupt to the microcontroller or in some cases certain registers are updated which trigger control operations to the microcontroller [13].

4. Implementation Of Ambient Light Sensors Using Intelligent Sensors Platform

Smart Sensors have a small memory and standard physical connections to enable communication with the Main Processor. Even though it has the advantages like reducing data communication with Main Processor for some default functions for a fixed time, low power consumption and easy integration. It has some glaring disadvantages like it is not possible to customize, data aggregation is not possible and calibration needs to be done by the Main processor which consumes additional power [8].

These drawbacks are overcome by the next generation of Smart sensors i.e. the Intelligent Sensors. The Intelligent sensors have the advantages of aggregating and processing the sensor data, and reconfigure the embedded functions according to the user requirements. This Intelligent Sensor Platform can be used to implement a Ambient Light Sensor with the advantage of implementing complex algorithms. Therefore the outcome is a combined architecture containing a microcontroller, Ambient Light Sensor and a small memory comprising the RAM, ROM and flash. The main feature of this Intelligent sensor platform is its ability to allow the user to customize the embedded logarithms that suits the user's needs and this can be achieved by reprogramming the flash on the fly. The basic structure of an intelligent sensor platform is shown below in Figure 10.

Figure 10:Basic Structure Of Intelligent Sensor Platform[8]

This platform has a low power microcontroller that aggregates the data from the Ambient Light Sensor and other related sensors and performs data processing and controls the LED backlight in this application. A Master Slave model is followed to enable the microcontroller to respond to the different sensors/interrupts [8].

Architecture:

The main constraints of this Intelligent Sensor platform is the size and power. The selection of a microcontroller is critical because it occupies higher power and size compared to the other components. The microcontroller chosen for this platform is the low power and compact Coldfire V1, it has a 32-bit RISC architecture. The Coldfire architecture is most suited for sensor applications as it contains a MAC unit which is capable of performing complex backlight control algorithms in a more efficient manner consuming less power. There are other several features such as timer, mailbox controller, PWM and GPIO (General Purpose Input Output).Interrupts and mailboxes that allows the microcontroller to switch between different sensors/ devices. The presence of a GPIO and PWM enables it to connect to different kind of sensors and devices. The 14 bit ADC is accessed through the GPIO and the digital data is processed by the microcontroller For communication with the Main processor and other external devices a SPI interface and I2C interface. Regarding the memory, the ROM memory contains basic functions in addition to the boot loader and the flash can be reprogrammed on the fly through I2C interface to suit the applications/user preferences. The memory available in this architecture is also used to store the processed data, which can be utilized by other sensors. This whole architecture can be realized in a die space of 3*3 mm package which is an essential requirement of smart phones [8] . The below Figure 11 shows the shows the microcontroller MMA9550L along with the other components.

Figure11:Intelligent Sensor comprising MA9550L,memory and other components [8]

The Intelligent Sensor Platform has several features which are listed below

1.Reduce data communication between the sensor hub and the applications processor 2.Less Power consumption, programmable sampling rate with sleep and wake up modes available 3. There is a continuous monitoring and calibration of the different sensors in the intelligent sensor platform 4. Less time for software development and each external sensor can be customized independently

The sensor data has been collected and processed by the microcontroller in the Intelligent Sensor platform, therefore this data can be directly send to the main processor (no data processing is done in main processor). In addition to this the stored signal data can also be used for the operation of other sensors such as a proximity sensor or combined with the other sensor data to perform a more complex control operation. The most important aspect of this platform is that all this data collection and data processing is done when the main processor is in sleep mode. So, when required this processed data is send in burst mode to the main processor through an interrupt thus saving more power. Experimental results have proven that a normal application processor like MX 51 consumes around 12mA current to perform backlight control based on ALS input, but this proposed architecture will be able to do the same control operation by consuming less than 100 µA as the main processor would be in sleep mode [8].

5. An ultra low power Ambient Light Sensor

In this approach an ultra-low power ALS has been designed to be implemented in Smart Phones . The Block diagram explaining the main components of this design is shown in Figure 12

Figure 12: Operation of the ultra low power ALS [10]

The photodiode senses the ambient light, the IR component of the ambient light is rejected due to the presence of IR rejection filter around the photodiode. The light sensed by the photodiode is converted to current and fed into an integrating type A/D convertor consisting of a current to frequency convertor and a 5 bits convertor. This A/D convertor receives the input current and transforms this analog signal into time (digital signal) using the current to frequency convertor. This convertor compares the input analog signal with a reference value which can be programmable. The time output from the convertor is counted using a 5-bits Counter and fed to the Data Register. The entire operation of the integrating A/D convertor is controlled by an oscillator clock and a control loop. When the value stored in the Data Register exceeds a threshold the microcontroller/LED driver does the operation of automatic dimming [10].

Integration time is the amount of time required by the integrating A/D convertor to sample the input analog signal and update the data register. Higher the resolution of the sensor more is the power consumption due to complex circuitry. Since the main use of ALS in smart phones is to reduce the power consumption without causing eye sore to the user, the resolution is set to a level of 5 in this model and the integration time is set to 400ms.Ttherefore the A/D convertor will sample the input light current every 4 seconds. With a resolution of 5 and integrating period of 400ms the sampling of the input photodiode current will be timed by 32 internal clock cycles. At the end of the integrating period the 5 bit counter will be reset. The control loop performs this operation of counting the internal clock cycles, resetting the 5- bit counter and updating the data register with the count value at the end of the integration period.

One of the major problems in circuits is the interference of noise signals, the Ambient Light sensors are not an exception to this. The light from artificial sources pick up periodic noises, the Integrating A/D convertor has excellent noise rejection ratio for periodic noises which are an integral multiple of its integration period. Thus by setting the integration period to an integer multiple of the periodic noise signals, then A/D convertor would effectively filter out these noise signals without the use of any additional circuitry thereby reducing power and cost. The performance results of this sensor is listed below in the Table 3

Type Of ALS

Digital

Integration period

4 s

Resolution

5

Range Of luminance

10 to 1000 lux

Power Consumption at Maximum Luminance

5 uA

Operating Voltage

5V

Table 3: Performance Summary [10]

8. Temperature Independent Low Power ALS

A serious problem encountered in many ALS and its related circuits when they are to be implemented in portable electronic devices such as Smart Phones is the effect of dark current and resolution of the conversion circuits. This model gives solution to these two major concerns in the design of an effective ALS

Dark Current Compensation :

Dark current in a photodiode increases with increase in temperature and thus causes a degrading effect on the ALS operation. In this model there are two photodiodes present PDlight which senses the ambient light and generates current I1 which include dark current and the other photodiode PDdark which generates the dark current I2 for compensation, the dark current is canceled out (I1 – I2) and thus only the ambient light current is used for A/D conversion and produces a appropriate digital signal for controlling the backlight. So, the effect of dark current is neglected to a great extent (around 95%) if magnitude and size of the dark current of both the photodiodes are equal [15].

Adaptive Resolution Adjustment:

The resolution of the ALS is directly related to the power consumption. Greater the resolution, greater is the power consumption and lower the resolution lower is the power consumption but that causes discomfort to the user. Therefore an adoptive resolution adjustment method is followed in this model so that the level of resolution changes according to the ambient environment lighting. In this mode the ALS can measure up to 10Klux. Therefore a 14 bit system is needed to measure this range of lux, by generating a count for every lux. In the adaptive resolution adjustment model a total of 10 bits is used rather than the usual 14 bits to change the resolution level and facilitate the efficient operation of the Ambient Light Sensor. The operating range is divided into 8 levels using 3 bits, but the number of fine steps in each level are same and determined by the counter but for each level the resolution i.e. the count/lux ratio is different. When the 7 -bit counter is full, it increments the 3 bit counter to extend the lux range the ALS could measure and when the 7-bit counter is half its capacity it decrements the 3 bit counter to reduce its range of lux measurement. For instance when the 3 bit counter has a value of "111" the counter (7-bit) is incremented for every 128 lux (Low Resolution) and if the 3 bit counter has a value of "000" the counter is incremented for every 1 lux (High resolution). This method also allows the ALS to dynamically change its range of measurement according to the environment. Therefore this proposed ALS with 10 bits occupies less power and size compared to the conventional 14 bit system. The power consumption of this ALS at its maximum luminance is 75 µA [15]. The below Table 4 lists out the lux range and lux/count ratio for each coarse level..

3- Bit Counter Value

Lux range

Lux/count

000

<128

1

001

<256

2

010

<512

4

011

<1024

8

100

<2048

16

101

<4096

32

110

<8192

64

111

<16384

128

Table 4: Lux/Count ratio adopted for the 8 levels [15]

V. Operation and ALS Algorithm Implementation in Current Ambient Light Sensors.

This section looks into the actual operation and various algorithms that have been adopted by the leading ALS manufactures to implement the ALS functionality. Four products have been discussed in detail under this section,

1. MAXIM MAX9635 Ambient Light Sensor The MAX9635 Ambient Light sensor is one of the most widely used ALS for many smart phones especially Samsung. It has a very low operating power and low operating voltage of 1.8V thereby enabling it for easy interaction with the microcontroller ports. This IC has an interrupt output pin that allows the main processor to stay in low power mode.

The preset registers in MAX9635 are the configuration register, interrupt enable register and Threshold Timer. The Microcontroller (master) has to write a bit into the IE register (slave) for enabling interrupt functionality. The master also writes a non -zero value into the Threshold Timer Register, so as to avoid any unnecessary ALS operation due to fleeting changes and this delay allows the brightness level to change smoothly rather than in an abrupt manner thus causing discomfort to the user. The user first reads the lux counts (ambient light sensed by the ALS) from the data registers LUX low byte and LUX high byte. Based on this information the current operating zone is found and the master (microcontroller) accordingly sets the Upper threshold register and Lower threshold register counts. For easy computation the lux values are converted into counts by dividing the lux values by 0.5 [16]. Due to the presence of a glass interface on the Smartphone display and small opening for the ALS to receive the light, the ambient light sensed by the ALS gets reduced by 5-10% this factor is taken into account while setting the threshold counts. The operating zones, their respective threshold counts and the requires backlight strength for each operating zone are listed below in the Table 5

Operating Zone

Backlight Strength

Upper Threshold Count

Lower Threshold Count

Dark

25 %

>10

<0

Dim

45 %

>50

<10

Home

65 %

>200

<50

Office

85 %

>200

<1000

Sunlight

100 %

>1000

<100000

Table 5: Operating zones along with their threshold and backlight strength (%) values [16]

When the ambient light is close to one of the operating zone's border then there would be fluctuations in the display brightness thereby causing discomfort to the user. To avoid this overlap zone is formed between the higher threshold count of one operating zone and the lower threshold count of another operating zone. Therefore a hysteresis is present which would avoid theses fluctuations. When an interrupt is caused, the microcontroller checks if the INTS bit is set in order to find out if the interrupt has been caused by MAX9635. If the interrupt has been caused by MAX9635 then it reads the ambient light and sets the threshold counts depending on the current operating zone. The algorithm has been illustrated in a vivid manner through the flow chat in Figure 13.

Figure 13: Pictorial Representation of the Algorithm [16]

2. TEXAS INSTRUMENTS LM3532

LM352 is a LED Driver with programmable Ambient Light sensing capability. This device is programmable over the I2C interface and has two external inputs for automatic brightness control of the display. The LM352's backlight current is controlled automatically based on the voltage at the two ALS inputs. LM352 has 32 internal resistors which are programmable at the two ALS inputs. There are a total of 5 ambient light brightness zones and there is a there is a total of 4 sets of Zone boundary registers. As the light sensed by the Ambient Light sensor increases there is a transition in the brightness zone which results in that particular Zone Boundary registers being enabled [17]. Each set of Zone boundary register consists of a Zone Boundary High Register and a Zone Boundary Low Register. The input to the 8-bit ADC is determined by the active ALS input (2-bit input) which is initially configured in the ALS Configuration Register as explained in Table 6

Active ALS input

Value

00

Max of ALS1 and ALS2

01

Min of ALS1 an ALS2

10

ALS1 only

11

ALS2 only

Table 6: Selection of Active ALS Input [17]

The ADC conversion usually takes 140 µs therefore once the digital conversion is complete the ADC Read back Register is updated with the value which is passed on as input to the LED driver to control the backlight. To avoid the change in brightness of the display due to any small fluctuations ALS averaging is done. There are a total of 8 average periods available for LM352 [17].

Algorithms:

LM352 utilizes two algorithms for effectively controlling the backlight, 1.Direct ALS 2.Down Delay. For both the algorithms certain condition have to be satisfied which are listed below [17]

1. Inorder to change the backlight current, the averager output should have either undergone a change (increase/decrease) for 3 consecutive periods or undergone a change in the one period and remain constant in the remaining two periods.

2 If condition 1 is satisfied and if the averager output shows a change in the same direction in the next period, then the backlight current is changed accordingly

3. If condition 1 is satisfied and in the next period the averager output remains the same then condition 1 has to be met again in order to change the backlight current

4. The backlight current changes its level only at the start of an average period

5. When the operating zone undergoes a transition the set of zone boundary registers are updated with the appropriate values in the same average period

Direct ALS Control:

In this part the backlight control by implementing Direct ALS control is explained with the help of the waveform in Figure 14. In this Figure the current average zone represents the current zone which is determined by the two ALS inputs and the average output corresponds to the brightness zone in the previous average period.

1. On startup the Averager Output is brought to the correct to the zone by the ALS fast startup.

2. In the 1st average period, the ALS voltage averages to zone 4

3. In the 2nd average period the ALS voltage drops down to zone 3 and stays in the same zone for the 3rd average period. The averager output shows a transition from zone 4 to zone 3 in the 3rd average period.

4. At the start of the 4th average period, the ALS voltage drops to zone 1 and the averager output stays put at zone 3.

5. During the 5th average period the ALS voltage drops to zone 0 and the averager output moves to zone 2. Since this the 3rd average period during which the averager output showed a change (decrease) and remained in the same zone, the backlight current is changed to suit zone 2.

6. In the 6th average period the ALS voltage increases to zone 3 and the averager output changes to zone 1. Since the change is in the same direction the backlight current again is reduced to suit zone 1 brightness target.

7. In the 7th average period there is no change in the ALS voltage but the averager output moves to zone 2, since the change is in opposite direction there is no change in the backlight.

8. In the 8th average period the averager output moves to zone 3.

9. In the 9th average period, the backlight current changes to zone 3 because this is the third period during which the averager output has shown an increase from zone 1 to zone 3.

10. During the next three average periods the averager output shows an increase and stays in the same zone for two consecutive periods, therefore at the start of the 14th average period the backlight current changes to zone 4 brightness target

Figure 14: Direct ALS [17]

Down Delay:

In the down delay algorithm additional delays are added into the operation of the ALS. The additional delays can be programmed by updating the ALS down Delay register. The register is provided with options to change the additional delays from 3 to 34 extra delays. In the waveform given below (Figure 15) the d own delay algorithm is enabled by updating the down delay register with 0x00 i.e. 3 additional delays therefore the averager output has to show a continuous change (increase/decrease) for 6 consecutive average periods or a change and remain in the same zone for 6 consecutive periods before forcing a change in the backlight current [17]. The main objective of implementing the down delay algorithm is to minimize power consumption to a great extent

Figure 15: Down Delay [17]

3. ADP5501 - Programmable Backlight Driver with Ambient Light Sensor Input

The ADP5501 is an LED Backlight driver which accepts the ambient light signal by means of an external sensor and performs effective control of the brightness of the Display in a Smartphone. The ADP5501 has an I2C interface for transmitting control signals/communication purposes and has the advantage of automatic dimming/lighting of the backlight, setting on/off timing between samples without any command signals from the Main processor, after it has been initially programmed. The ADP5501 works over a range of three operating levels, these levels are set based on the value of the bits BL_LVL in the configuration register [18]. These levels along with the bits and boundary registers are listed below in the Table 7.

Backlight Operating level

BL_LVL

Target Boundary registers

Daylight

00

DAYLIGHT_MAX

DAYLIGHT_MIN

Office

01

OFFICE_MAX

OFFICE_MIN

Dark

10

DARK_MAX

DARK_MIN

Table 7: Selection Of Backlight Operating Level based on BL_LVL bit [18]

The ambient light is sensed by the external photo sensor connected to ASPS501 and the output is fed into the 16-bit ADC. The digitized output is fed to two comparators L2 and L3 (representing the backlight operating levels Office and Dark respectively). When the ADC output falls below the trip point set for the comparator L2, an interrupt is triggered and the backlight brightness is changed to Office Mode and L2_OUT bit is set indicating that the current operating level is Office. The same operation takes place in the comparator L3 by setting L3_OUT. If then both the comparators do not set their output bits (L2_OUT and L3_OUT) it indicates that the current backlight operating level is Daylight and the corresponding backlight strength is set. The main advantage of this chip is its feature to distinguish between three operating levels based on two comparisons thereby leading to reduced power consumption [18]. The ADP5501 using the internal state machine automatically controls these transitions between different operating levels by enabling the BL_AUTO_ADJ bit (Backlight Level Auto Adjustment Bit). Table 8 below illustrates how the operating zone is determined based on the bits BL_AUTO_ADJ, L3_OUT and L2_OUT.

BL_AUTO_ADJ

L2_OUT

L3_OUT

Backlight Operating Level

0

X

X

Set by user

1

0

0

Daylight

1

0

1

Office

1

1

0

Dark

Table 8: Selection Of Operating Level based on the control bits

4. NO A3301 -Digital Proximity Sensor with Ambient Light Sensor:

The NO A3301 combines of the features of both the Proximity Sensor and the Ambient Light Sensor into a single package with the powerful Interrupt and low power consumption. The ALS in this package has a dynamic range as low as 0.05 lux to as high as 52 Klux with a 21 bit effective resolution, hence it exactly mimics the human eye. For enabling low power consumption it has an option for Automatic Power Down after a single measurement [19]. The ALS senses the ambient light and based on this input the 16 bit ADC produces a proportional digital out over the I2C interface to be stored in the ALS_DATA Register based on which the display brightness is adjusted. Table 9 lists out the different data registers and bits necessary for ALS operation.

Register / bit

Operation

ALS_TH Resisters

Upper and Lower Threshold settings

ALS_Config Registers

Includes configuration for integration time, enabling Hysteresis and triggering Hysteresis

ALS_Control Register

Select between continuous measurements or single shot mode

ALS_Interval Register

The time between two measurements/sampling is set

ALS_DATA Register

Store the 16-bit ADC Output

ALS_Repeat bit

Initiates new measurements based on values set in ALS_Interval register

ALS_Oneshot bit

Triggers the ALS operation, it is cleared after the sampling operation

Table 9: Data Registers/bits available in NO A3301[19]

The ALS in NO A3301 has an option to select between One shot mode and continuous measurement. In One shot mode the ALS completes its operation, clears the ALS_OneShot bit and stays in power down mode waiting for the next command signal. But in Continuous Mode the ALS operation is repeated at specified intervals set in the ALS_Interval Register. Upon ALS initialization the ALS_Config Register is configured and the photodiode senses the ambient Light, passes it on the ADC which stores the corresponding digital signal in the ALS_DATA Register. Depending on the Mode selected the backlight strength is controlled and the ALS goes into power down mode(One Shot mode) or samples the next signal based on the interval set (Continuous Measurement) [19]. The time between two integrations in an ADC depends on the interval set in the ALS_Interval Register and the Integration Time in the Configuration Register. The different scenarios are listed below in Table 10.

Scenario

Time between two Integrations

Power

Interval = 0

5 us

ON

Interval <=Integration period

10ms

ON

Interval>Integration period

Set Interval

DOWN

Table 10: Selection Of Interval between two samples [19]

VI. Alternatives To Ambient Light Sensors:

Even though there have been many improvements made in the field of Ambient Light Sensors, certain other new technologies have emerged as a replacement for Ambient Light Sensors in order to achieve much lesser power consumption, better efficiency and faster operation. Since these technologies are still under research the results may not be as high as ALS but in future these methodologies might work better the current Ambient Light Sensors. Two such novel methodologies have been discussed below

1.Automatic Brightness Control Of Smartphone in Low Illumination

The main requisite for this method is the presence of a front panel camera. In low ambient brightness, the camera on the front panel captures the image of the contrast ratio between the user's face and the background is computed using a face detection technique. If the user's face is dim compared to the background then the control software would increase the brightness of the display. But if the use's face image is more brighter than the background then the brightness would be lowered in order to avoid glaring effect and better readability.

The user's face from the image captured by the front panel camera is calculated/detected using a face detection technique proposed by Soriano. This face detection technique has been adopted because a simple color model is used and it is easier to implement and perform calculations in a shorter span of time. In this technique the user's skin area in the image is bounded on a normalized r-g plane, the two equations listed below calculate are used to detect if a particular pixel refers to the skin [9].

r = R/(R+G+B)

g = G/(R+G+B)

If Q+ < g < Q- then the pixel is marked as skin. After the detection of the user face and the background from the image the average luminance of each pixel is computed using the below equations [9] Q+ = -1.3767r2 + 1.0743 r + 0.30

Q- = -0.7760 r2 + 0.5601 r + 0.10.

The output of face detection technique for an image is represented in Figure 16. Figure 16: Separation of face and Background in an image [9]

After the face and background have been separated from the image their contrast ratio is calculated. This contrast ratio cannot be directly used for brightness compensation as it would cause discomfort to the user, therefore a Negative feedback system is implemented as shown in Figure 17

Figure 17: Negative Feedback System for setting Brightness Compensation [9]

In the Feedback Control Loop represented in the above system, the following steps are performed

1. The image of the user is captured by the front panel camera.

2. The Control software and face detection technique identifies the face and the background and face to background contrast ration is set.

3. The compensation value is computed, if this value does not exceed the set threshold, the brightness is not changed, but if the threshold value is exceeded then the brightness compensation value is increased/decreased in small steps to avoid any discomfort to the user [9].

This approach has been implemented as an application Samsung Galaxy S2. The user is provided with options to set his desired face to background ratio, brightness values or else the default values will be used. It has been observed that the current technique works well only for images with a darker/dim background, hence future work is being carried on to inculpate this technique even for images with brighter background. This approach works well even if the Smartphone is held at a farer distance from the user as in such a scenario the screen becomes brighter.

2.Content Based Adaptive Control (CBAC)

In Content Based Adaptive Control, the Backlight brightness is changed according to the image. For instance for displaying a dark image usually the filter blocks majority of the backlight hence causing the image to appear dark, instead the Backlight brightness can be reduced and the filter can be made to allow more light thereby boosting the image. The image quality perceived in both the cases is almost same but in the second case the backlight has been dimmed to achieve the same image quality as shown in Figure 18. Histogram Analysis is done in order to find the maximum brightness value for the image. Based on this value, the image is boosted [20]. There are several types of backlight control possible, in this section we will be discussing about 0-dimensional control In 0 - dimensional control the lighting of the entire image is changed uniformly and smoothly.

Figure:18 Content Based Adaptive Control [20]

The two major constraints in Content Based Adaptive Control are "Washout" and "Flickering". When the backlight is reduced the dynamic range is also reduced, therefore pixels requiring a higher range of brightness would be cut out, this is termed as "washout". When the target brightness value varies by a great extent between two frames, changing the brightness value by such a great amount causes discomfort to the user, this is termed as "Flickering".

In order to eliminate these constraints the histogram analysis is done for each color component and their brightness value is computed. The maximum brightness among these color components is taken as the target brightness. "Washout" can be either eliminated by clipping of those pixels which cannot be perceived by human eye or enabling a smooth roll of to image transform which ensures there is no clipping of any color components [20]. If the brightness of one frame is very much greater than the succeeding frame then the target brightness value would be the average of the brightness values of the two frames. This approach eliminates "Flickering"

There are several Content Based Adaptive Control techniques adopted by several manufacturers, these techniques reduces the power consumption by around 50 % without distorting the image quality.

VII. Conclusion

Reducing the power consumption of the LCD/TFT Displays in Smartphones is a very critical requirement in during the design of a Smart Phone. In this paper the various ALS topologies and the various new methods/prototypes of implementing the Ambient Light Sensors in Smartphones has been discusses in detailed form a hardware and software perspective. This paper has also elaborately discussed on how the ALS operate in the smartphones available in the market along with their algorithms. Overall it can be summarized that the manufacturers have realized the need to reduce power consumption in displays, therefore the standards of the Ambient Light Sensors would see an upward trend in the days to come and few alternatives such as face detection and Content Based Adaptive Control will also be explored in depth.