Joaquin Font

Estimates provided by the Insurance Institute for Highway Safety show that over 1.5 million deer-vehicle collisions occur per year in the United States, causing 150 occupant deaths, and over $1 billion in vehicle damage. While a substantial body of literature addresses deer behavior and infrastructure-related factors in deer collisions, on patterns in deer-vehicle collisions, and on human factors influencing the frequency of collisions, few studies have addressed how to accurately simulate intelligent vehicles interacting with animals on a roadway. Fewer still address collision avoidance system design with respect to animal-vehicle collisions specifically.

 

The present study uses a simple, flexible model for deer motion during roadway crossing, and attempts to find particularly dangerous combinations of deer model parameters using a Genetic Algorithm (GA), identifying deer crossing events with heightened risk to vehicle safety. Then, using a simulated autonomous vehicle equipped with model-predictive control (MPC) for collision avoidance, and traveling at the posted speed limit for the road, a GA is iterated through a sufficient number of generations to find the “worst-case deer” for a given vehicle speed and road layout. If collisions are probable after training, the vehicle’s forward speed is reduced until a speed resulting in an acceptable collision risk is found. This process is performed for a series of locations along the roadway throughout the high-risk area. The overall objective is hence to find the maximum safe speed for an autonomous car at a specific point within a given road to safely avoid a collision with a deer appearing at a consistent distance ahead of the vehicle.

Joaquin Font is a senior mechanical engineering and economics dual degree student at Lafayette College. Born in Buenos Aires, Argentina, Joaquin is passionate about the automotive industry, and has been performing research in vehicle dynamics under the supervision of Prof. Alexander A. Brown for over two years. He is currently working on his honors thesis titled “Determining map-referenced maximum safe speeds in high-risk deer-vehicle collision areas using autonomous vehicle simulations," through which he aims to improve roadway safety by designing a new way to calculate speed limits in areas where deer-vehicle collisions are highly occurring. Joaquin is also a member of the American Society of Mechanical Engineering (ASME), and served as student chapter president for the 2017-18 academic year. Outside of engineering, he enjoys travelling, photography, and running.