This thesis addresses a proposition for a smart traffic light design with the use of radar. We focused on a pedestrian traffic light, placed near continuous traffic, which presently only allows street crossing if pedestrians express this intention by pressing a nearby button. Our design proposes contactless recognition of such intent and following change in traffic signals, thus allowing street crossing for pedestrians.
For pedestrian recognition, we are using an FMCW radar, which, in comparison to other similar radar solutions, precisely determines the position, speed, acceleration and direction of the observed target. FMCW radar achieves this precision with the constant periodic transmission of a signal called chirp.
We used various methods to test radars' capabilities and made several square corner radar trihedrals for this purpose. From the test results, we determined the radar's operating parameters for tracking pedestrians. For tracking pedestrians, we use the GTRACK algorithm, modified to run on an external computer or processor to reduce the load on the CPU, which is integrated inside the radar.
For traffic light activation, we designed an algorithm that replaces the button usually placed next to the traffic light. The primary purpose of this algorithm is to correctly identify pedestrians who intend to cross the street. The algorithm has to ignore other objects or pedestrians who do not intend to cross the street and are merely passing the street crossing.
|
