Dedicated Short Range Communications (DSRC) are known as the main option to be mandated by the United States Department of Transportation for Vehicle-to-Vehicle (V2V) communication. If so, the industry has already demonstrated DSRC-enabled phones at no additional cost. Therefore, every smartphone can be used to protect its owner while outdoor. The objective of this project is to propose a context-aware algorithm to cope with the noisy GPS sensors in the urban canyons to classify the pedestrians to two groups of Vulnerable Road Users (VRU), i.e. pedestrian crossing the street, and the ones walking on safer areas, i.e. walking on the sidewalk. This classification could be further used to assign different transmission parameters to these two groups, e.g. lower/higher inter-transmission interval to keep the channel load low and reduce the unnecessary transmissions.
Communication-based Vulnerable Road User (VRU) safety system frequently broadcasts every user's position information, e.g. 5 times per second. Given the cars are able to overhear this information, they are able to track the pedestrians to avoid colliding with them. In crowded areas, however, number of transmitters, i.e. DSRC-enabled smartphones, could be more than what the wireless channel capacity can tolerate. Therefore, using context-aware algorithms, the aim is to cut the unnecessary Personal Safety Messages (PSM) to reduce the channel load and interference on the channel. GPS inaccuracy in areas with many sky scrapers is challenging the idea. In this work, a heuristic is introduced to overcome this issue and identify the pedestrian walking on the road with a high confidence.
The key idea of the proposed CTP is to track multiple context clues that indicate that a smartphone user is not currently a vulnerable road user and to reduce or eliminate personal safety message transmission in this case. In particular, the design focuses on the key challenge of identifying the many smartphone users who are in relatively safe location on sidewalks or in pedestrian zones even when the positioning data available to the smartphone is affected by errors on the order of tens of meters, as frequently the case in urban canyons. It accomplishes this through a map of common street crossing points, where pedestrians walk onto the street, and an adaptive guard zone around these crossing points that is adjusted based on the positioning error estimate. The proposed algorithm is evaluated through extensive simulations for the Manhattan area in New York city.
The Information Age reflects how fresh the pedestrian's information is at the receiver. The information age is the time since the data sample, i.e. the GPS sensor data in this case, in the last received message was generated on the transmitter side. The following figure shows the age of information. The Information Age is sampled every 10 msec and the calculation is limited to the cases where the transmitter is a VRU in the street and is less than 150m away from the receiver. The observation is that with Oracle solution, i.e. the best possible results based on the previous assumption in the current setup, about 90% of age samples are less than 440 ms. However, baseline 5Hz algorithm provides 1700 msec for the same criteria. As our CTP solutions, for different configurations of the proposed algorithm, CTP 90% of samples are less than 710 msec for the best configuration.