Mobile Sensing Revolution Keynote Past Times Hari Balakrishnan
Last calendar week I was inwards Boston to attend International conference on Parallel in addition to Distributed Processing Systems (IPDPS), in addition to I volition postal service some of my notes from this conference here.
Coincidentally, Distributed Computing inwards Sensor Systems (DCOSS) was also existence held at Boston at the same dates alongside IPDPS, in addition to I could solely reserve my room at the DCOSS hotel instead of the IPDPS hotel. Since the DCOSS keynote on mobile sensing past times Hari Balakrishnan sounded to a greater extent than interesting to me than the IPDPS starting fourth dimension twenty-four lx minutes menstruation keynote, I decided to sneak inwards to the DCOSS room in addition to attend that keynote instead. Hari's utter had ii parts. He spent most of his fourth dimension on the starting fourth dimension part, in addition to had to rush through the 2d part.
Smartphones are genuinely mobile devices, in addition to equally a number they deed through dissimilar environments in addition to sense dissimilar network conditions. Even during a piece of cake walk, signal-to-noise-ratio (SNR) changes drastically alongside time, in addition to the bit-error-rate (BER) varies 5-6 orders of magnitude inside this fourth dimension interval. So at that spot is a huge volatility of the network connectedness stemming from mobility. The query is: How exercise nosotros bargain alongside these rapidly changing networking conditions?
This is non a novel problem. The means 802.11 deals alongside this is past times using chip charge per unit of measurement adaptation. To this end, 802.11 measures the wireless channel at the device (the package loss rate, SNR, or BER), in addition to determines the best chip charge per unit of measurement in addition to switches to that at the wireless access indicate (WAP) for communicating to that device. But the work alongside this is, for smartphones, which are genuinely mobile devices, these measurements are rapidly outdated.
So, how exercise nosotros stair out in addition to determine the best chip charge per unit of measurement for smartphones? Hari's stance is to move sensors on the smartphone to solve/alleviate this problem. Smartphone has several rattling adept sensors including accelerometer, compass, gyroscope, GPS, lite sensor, microphone, in addition to fifty-fifty a barometer. We tin move the readings from these sensors to say which means the device is oriented, its position, direction, in addition to speed, in addition to move these hints to adjust to the dissimilar mobility modes accordingly. For example, the accelerometer tin respect walking in addition to this tin aid improve charge per unit of measurement adaptation at the MAC layer. The compass tin respect heading in addition to this tin aid improve AP association at the network layer. The gps tin respect speed in addition to this tin aid improve vehicular routing at the network level. The smartphone tin also move hints (sensor readings) from neighboring smartphones equally business office of a mobility hint protocol. You tin also easily construct a crowdsourced network quality/connection monitoring service.
Hari's grouping focused on detecting motion in addition to built the "rapidsample" service for quick motion detection on the Android. Using rapidsample your device tin deed into an aggressive means upon detection of motion inwards club to address/alleviate the movement-induced network problems. Rapidsample tin reliably respect motion inside 10-100ms on commodity Android devices past times considering accelaration, magnitude, in addition to variance. Rapidsample is able to larn pretty low-energy sensor readings from the accelerometer on the Android. When I mentioned to Hari about our findings almost costly accelerometers on Android, he said that this toll is due to the inefficient Java implementation of the drivers on the Android. Hari's grouping has custom implemented their accelerometer sensing inwards C to attain energy-efficiency in addition to depression latency.
Using the sensor-augmented protocols (by employing rapidsample), they were able to attain 40-60% higher throughput for a 50% mobile user. And for the access indicate choice work for android for VOIP, they were able to attain 40% fewer handoff disruptions without whatever modify on the WAP part. Finally, they were able to attain 20 times reduction inwards probe bandwidth equally good cheers to rapidsample. Some limitations of their electrical current piece of work are that 1) spell accelerometer in addition to compass are inexpensive to move GPS is not, in addition to 2) calibration across dissimilar device types is needed (while the motion hint required no calibration, the walking hint did).
Hari is working alongside insurance companies for assessing driver safety, past times computing driver endangerment alongside existent information from smartphone apps. Smartphones are substitution to volume adoption of this idea, then they built an Android background app for this. Their app is on the Android app shop for people to try.
Their app uses solely WAP based localization to proceed a low-energy profile. As a number they claim that the app running inwards the background tin bridge upto 50 hours of battery lifetime. The work is that the WAP signals tin endure pretty sparse, then they had to prepare a trajectory mapping algorithm that employs a hidden markov model to move both previous in addition to hereafter place information from the users. This trajectory mapping algorithm plant on the cloud in addition to tin endure pretty computationally intensive. For event a Houston-Chicago effort had recorded solely xl WAPs along the means in addition to the trajectory mapping algorithm took 1 lx minutes to compute the trajectory taken past times the car.
Their app collected 2 1000000 full miles inwards tests then far. They flora that user mileage is rattling important, in addition to sure enough combinations of accelerometer in addition to speed is problematic, which they capture equally a colored endangerment There is no basis truth to claim that the app makes the drivers safer drivers, precisely Hari claims that looking to the accelerometer logs you lot tin run across ameliorate driving habits (since the drivers are incentivized to larn a discount on their insurance premium).
They flora that most of the users don't attention to plug inwards their smartphones to the automobile for recharging equally they run across this equally a burden. So Hari's number 1 wishing for the hereafter smartphones is to include lower-power (virtually zero-power) sensors. This tin endure done past times a ameliorate pattern of the hardware in addition to software. As Hari puts it, you lot desire the sensors to run continuously in addition to lay the sensor information into a buffer in addition to when the CPU wakes upwardly it tin larn the information from the buffer.
Coincidentally, Distributed Computing inwards Sensor Systems (DCOSS) was also existence held at Boston at the same dates alongside IPDPS, in addition to I could solely reserve my room at the DCOSS hotel instead of the IPDPS hotel. Since the DCOSS keynote on mobile sensing past times Hari Balakrishnan sounded to a greater extent than interesting to me than the IPDPS starting fourth dimension twenty-four lx minutes menstruation keynote, I decided to sneak inwards to the DCOSS room in addition to attend that keynote instead. Hari's utter had ii parts. He spent most of his fourth dimension on the starting fourth dimension part, in addition to had to rush through the 2d part.
How tin sensors improve networks?
The crossover happened this year: smartphone & tablets accessing the Internet overtook PCs & laptops accessing the Internet. Smartphone adoption charge per unit of measurement is vii times the charge per unit of measurement of increase of basis population, in addition to at that spot are to a greater extent than people using smartphones than people that guide keep access to electrics, in-house toilet, or toothbrush.Smartphones are genuinely mobile devices, in addition to equally a number they deed through dissimilar environments in addition to sense dissimilar network conditions. Even during a piece of cake walk, signal-to-noise-ratio (SNR) changes drastically alongside time, in addition to the bit-error-rate (BER) varies 5-6 orders of magnitude inside this fourth dimension interval. So at that spot is a huge volatility of the network connectedness stemming from mobility. The query is: How exercise nosotros bargain alongside these rapidly changing networking conditions?
This is non a novel problem. The means 802.11 deals alongside this is past times using chip charge per unit of measurement adaptation. To this end, 802.11 measures the wireless channel at the device (the package loss rate, SNR, or BER), in addition to determines the best chip charge per unit of measurement in addition to switches to that at the wireless access indicate (WAP) for communicating to that device. But the work alongside this is, for smartphones, which are genuinely mobile devices, these measurements are rapidly outdated.
So, how exercise nosotros stair out in addition to determine the best chip charge per unit of measurement for smartphones? Hari's stance is to move sensors on the smartphone to solve/alleviate this problem. Smartphone has several rattling adept sensors including accelerometer, compass, gyroscope, GPS, lite sensor, microphone, in addition to fifty-fifty a barometer. We tin move the readings from these sensors to say which means the device is oriented, its position, direction, in addition to speed, in addition to move these hints to adjust to the dissimilar mobility modes accordingly. For example, the accelerometer tin respect walking in addition to this tin aid improve charge per unit of measurement adaptation at the MAC layer. The compass tin respect heading in addition to this tin aid improve AP association at the network layer. The gps tin respect speed in addition to this tin aid improve vehicular routing at the network level. The smartphone tin also move hints (sensor readings) from neighboring smartphones equally business office of a mobility hint protocol. You tin also easily construct a crowdsourced network quality/connection monitoring service.
Hari's grouping focused on detecting motion in addition to built the "rapidsample" service for quick motion detection on the Android. Using rapidsample your device tin deed into an aggressive means upon detection of motion inwards club to address/alleviate the movement-induced network problems. Rapidsample tin reliably respect motion inside 10-100ms on commodity Android devices past times considering accelaration, magnitude, in addition to variance. Rapidsample is able to larn pretty low-energy sensor readings from the accelerometer on the Android. When I mentioned to Hari about our findings almost costly accelerometers on Android, he said that this toll is due to the inefficient Java implementation of the drivers on the Android. Hari's grouping has custom implemented their accelerometer sensing inwards C to attain energy-efficiency in addition to depression latency.
Using the sensor-augmented protocols (by employing rapidsample), they were able to attain 40-60% higher throughput for a 50% mobile user. And for the access indicate choice work for android for VOIP, they were able to attain 40% fewer handoff disruptions without whatever modify on the WAP part. Finally, they were able to attain 20 times reduction inwards probe bandwidth equally good cheers to rapidsample. Some limitations of their electrical current piece of work are that 1) spell accelerometer in addition to compass are inexpensive to move GPS is not, in addition to 2) calibration across dissimilar device types is needed (while the motion hint required no calibration, the walking hint did).
Mobile sensing alongside driver safety
Today nosotros however human face upwardly severe route shipping challenges. There are 10 1000000 accident/year inwards USA which results inwards 35K deaths annually. Insurance companies offering incentives (upto 15% of the premium) via usage-based insurance (UBI) for safer driving. But UBI is implemented today using embedded devices which makes for a hard deployment in addition to maintenance.Hari is working alongside insurance companies for assessing driver safety, past times computing driver endangerment alongside existent information from smartphone apps. Smartphones are substitution to volume adoption of this idea, then they built an Android background app for this. Their app is on the Android app shop for people to try.
Their app uses solely WAP based localization to proceed a low-energy profile. As a number they claim that the app running inwards the background tin bridge upto 50 hours of battery lifetime. The work is that the WAP signals tin endure pretty sparse, then they had to prepare a trajectory mapping algorithm that employs a hidden markov model to move both previous in addition to hereafter place information from the users. This trajectory mapping algorithm plant on the cloud in addition to tin endure pretty computationally intensive. For event a Houston-Chicago effort had recorded solely xl WAPs along the means in addition to the trajectory mapping algorithm took 1 lx minutes to compute the trajectory taken past times the car.
Their app collected 2 1000000 full miles inwards tests then far. They flora that user mileage is rattling important, in addition to sure enough combinations of accelerometer in addition to speed is problematic, which they capture equally a colored endangerment There is no basis truth to claim that the app makes the drivers safer drivers, precisely Hari claims that looking to the accelerometer logs you lot tin run across ameliorate driving habits (since the drivers are incentivized to larn a discount on their insurance premium).
They flora that most of the users don't attention to plug inwards their smartphones to the automobile for recharging equally they run across this equally a burden. So Hari's number 1 wishing for the hereafter smartphones is to include lower-power (virtually zero-power) sensors. This tin endure done past times a ameliorate pattern of the hardware in addition to software. As Hari puts it, you lot desire the sensors to run continuously in addition to lay the sensor information into a buffer in addition to when the CPU wakes upwardly it tin larn the information from the buffer.
0 Response to "Mobile Sensing Revolution Keynote Past Times Hari Balakrishnan"
Post a Comment