Driver Classification

This student team worked to add a new classification layer and behavioral prediction to autonomous driving. The main objective was to classify and predict reckless drivers by identifying specific trends and indicators while continuously monitoring the vehicle's surrounding area. Due to the complexity of the objective, the project scope was divided into multiple phases, with Phase I taking place in the 2022/23 Industry Capstone Program. PHASE I: This phase was meant as a proof of concept. The algorithms the student team worked to create in this phase are intended to be tested on a small-scale database. A small-scale database contains a limited number of training data and testing objects, which was to be defined in detail during the early stages of the project. The Phase I student team was expected to: - identify the specific indicators for reckless driving - generate a prediction algorithm based on the identified indicators - detect and classify a specific vehicle - track and monitor the behavior of the classified vehicle. Phase I and 2 student teams worked to classify and identify driving characteristics that represent reckless behavior, including creating requirements that would identify reckless behavior. For these phases and subsequent phases, a stretch goal was for the student team(s) to identify additional driving characteristics/targeted behavioral patterns, such as: “dead pedal”, speeder, weaver, angry/aggressive, brake checker, vehicle speed limited, strict speed limit follower, normal, erratic, inattentive, defensive, etc., and define written requirements for these additional driving characteristics/targeted behavioral patterns.

Latest news