GSoC’21 RoboComp project: Software agent for traffic and pedestrian traffic monitoring in outdoor environments using RGB cameras

19th June, 2021

About me

I’m Amarthya Sasi Kiran, a 4th year undergraduate, pursuing my B.Tech + MS by research dual degree at IIIT Hyderabad, India. Following my passion in the area of Robotics, I devote my time and efforts in this direction. My research interests are in the area of Robotics. Autonomous Vehicle Navigation and Behavioral motion and path planning are my current areas of research in particular, at the Robotics Research Centre(RRC), IIIT Hyderabad, India.

About the Project

In today’s world, a lot of educated decisions are made using the information collected by robots and are aided by the inferences given by the software agents which make use of the information collected by the other machines and robots. One such crucial field, where the software agents can contribute is “Pedestrian and Vehicle Traffic Monitoring”, where they can help us identify the hotspots and conflict zones like Pedestrian Crossings, Traffic Signals, Traffic jams, etc. The aim of this project will be to achieve this by using detection networks and depth networks. The depth information provided is exploited in tandem with the pixel locations of the points of interest provided by the detection network to get the information on the 3D position of the Vehicles and Pedestrians w.r.t to the RGB camera. Through this, the information provided by machines like RGB cameras can be exploited maximally to make better decisions based on the density of people and vehicles present in an area.