andreavaccari.com

Today in a MIT CSAIL seminar, Dr. Ramachandran Ramjee presented his ongoing research on smartphone-based monitoring of traffic and roads conditions. The research is motivated by a need to go beyond GPS-based monitoring which proved to be ineffective in complex environments like those of developing countries. These environments (R.R. showed a video of an intersection in Bangalore, India) are characterized by an heterogeneous mix of vehicles, fatigued road pavement, liberal honking, continuous stop-and-go traffic, and so fore.

Ramachandran first talked about the benefits of using smartphones over GPS-based and infrastructure-based systems (proliferation, programmability, …). He also talked about the limitations of such devices, and in particular the energetic constraints that they have. A idle smartphones consumes 183 mW, where an embedded WiFi receiver consumes alone 772 mW, a GPS receiver consumes 617 mW, and a microphone consumes 223 mW. Luckily other sensors require less power: an accelerometer only needs 2 mW and a bluetooth receiver around 20 mW. For this reason, R.R. suggested that it is necessary to use only the latter as much as possible, and to activate the former only when needed.

In particular, the accelerometer is used for the detection of brake stops and bumps or potholes, which then triggers an approximate computation of the position of the vehicle based on cellphone towers triangulation. If the server that collects the data receives many reports from the same area, it can send a signal back to the smartphone requiring the activation of the GPS receiver to get, if possible, a better estimation of the position.

The microphone is instead used to detect the type of vehicles (exposed or enclosed) according to the ambient noise, and also to detect honks which can be used to understand the traffic load and entropy. The bluetooth, finally, can be interestingly used to understand if the user is moving on a private or public vehicle, where the latter presents a more stable bluetooth neighbourhood. All these sensors and the information that they provide can then be fused together in an effort to build a better understanding of the traffic and roads conditions.