Adaptive Motion Planning - Research Group
Home Research People Workshops Outreach Resources Videos    

Picture+ Hao-Tien (Lewis) Chiang
PhD Student

Adaptive Motion Planning Research Group
Department of Computer Science
Mailing address:
Mail stop: MSC01 1100
1 University of New Mexico
University of New Mexico
Albuquerque, NM 87131-0001
office: 3400A Farris
telephone: 505-277-8912
fax: 505.277.6927

I got my BS degree in Atmospheric Sciences and Physics at the National Taiwan University. He first joined the University of New Mexico as a PhD student in Physics and worked on developing Quantum Algorithms for 3 years. I then transfered to the PhD program in Computer Science at UNM in Spring 2015 under the supervision of professor Lydia Tapia.

Since joining my PhD program in January 2015, I've been developing methods for planning in dynamic environments. In one such method, Monte Carlo simulations were used to predict the stochastic motion of obstacles as well as robot sensor noise. These predictions are used to build trees of possible robot motion in order to identify collision-free paths [Chiang et. al. IROS 17, Chiang et. al. IROS 16, Chiang et. al. IROS 15]. In another example, we utilized stochastic reachability analysis, which is a formal method that provides navigation safety guarantees even in the presence of stochastic obstacle motion. However, since the computational cost of this method scales exponentially with the number of obstacles, we combined it with an artificial potential field-based approach in order to avoid collisions in real-time with up to 900 moving obstacles [Malone et. al. T-RO 17, Chiang et. al. WAFR 14, Chiang et. al. ICRA 15]. We also developed a reinforcement learning-based method where the agent is trained to avoid obstacles in a simple environment. The learning result can be transferred to avoid collisions in large complex environments with hundreds of stochastically moving obstacles [Faust et. al. ICRA 16].

CV:   pdf

 Projects: Moving Obstacle Avoidance |  Adaptive Quadrotor Control |  Human-Automation Collaboration