CityDrive: A Map-Generating and Speed-Optimizing Driving System
Absent of traffic light information, drivers can hardly adopt an appropriate speed and thus they often come to a halt reaching intersections. This stop-and-go driving pattern causes increased fuel consumption, air pollution, traffic congestions and road accidents.
There have been many traffic light control systems around the globe, but the high cost of infrastructure and maintenance hinders their wide deployment. However, speed-advisory systems enabled by on-vehicle devices are much cheaper and easier to deploy.
The first challenge of such systems is to get the traffic signal schedule in complex intersections. The second challenge is to get map information and calculate the distance.
Facing these challenges we devise and implement a speed-advisory driving system called CityDrive, which harnesses the sensor and GPS data from a wide participation of smartphones to suggest proper speed for drivers so that they arrive at inter-sections in green phase. CityDrive first generates a road map and then infers traffic signal schedules, using only smartphones and a server. CityDrive does not eliminate stops at intersections, but it tries to maximize the probability that vehicles cruise through intersections in green phase. Both simulation and real test show that this continuous speed advisory service effectively can smooth traffic flow and significantly reduces energy consumption.
We explore the accelerometer and magnetometer on android phones to detect vehicles’ acceleration events. This figure shows the acceleration magnitude along the direction the vehicle’s heading. Acceleration from zero speed is an indicator of the existence of an intersection. And we also use that data to infer traffic signal schedules.
To locate candidate intersections, we apply a method similar to mean shift, in which successive computations of the mean shift yield a path leading to a local acceleration density maximum
The figure shows the process to locate candidate intersections. The magenta arrows are acceleration vectors. The green circles are the shifting circles, and the red ones are final circles (candidate intersections).
To generate road segments, we use second order B-spline curves generated by consecutive anchor points. The anchor points are points that represent the mean position and direction of collected road GPS data. This is the generated road segments and the real map in Google Map.
Traffic signal inference:
To infer the traffic signal schedule in complex intersections with multiple sets of traffic lights, we first give each in-coming and out-going branch a index number. Then we regard each traffic movement from one in-coming branch to one out-going branch as a state. We use vehicles’ acceleration time to deduce the traffic signal cycle length and link multiple states into loop chains. Then using the acceleration events sent to the central server, our system employs a method to minimize the mean square error of the predicted time and real event time. In this way we can utilize future acceleration events to calibrate the traffic signal schedule.
Real world experiments:
We did our real world test in Zizhu Hi-Tech Industrial Development Zone, Shanghai, China. The speed recordings show that the vehicle’s speed is significantly smoothed. This can be shown in the speed-acceleration histogram below.
Further calculation shows that our system yields a surprising 58.8% saving of kinetic energy, which indicates that this intelligent transportation system has a promising role in relieving our energy crisis and environmental pollution.