Register now After registration you will be able to apply for this opportunity online.
Collision Avoidance Motion Planning for Robotic Manipulation
Obstacle avoidance is a fundamental problem in robotics and is particularly relevant for field robots operating in cluttered environments, e.g. forestry robots. In this project, we would like to develop a real-time feedback framework for this task in parts: perception and detection for obstacles with onboard sensing, and planning for robotic manipulator collision avoidance in cluttered environments
Keywords: Perception, Sampling and Optimization Based Planning, Robotic Manipulation
Obstacle avoidance is a fundamental problem in robotics and is particularly relevant for field robots operating in cluttered environments, e.g. forestry robots. While sampling- and optimization-based methods are widely used for current robotic applications, they are not always able to handle both kinodynamic considerations and arbitrary-shaped obstacles in real-time. Previous research at RSL [1] combined the advantage of both classes of algorithms to achieve efficient and feasible motion planning. Recent advances in MPC algorithms also [2] showed the possibility to deal with nondifferentiable polytopic constraints using dualization. However, the efficiency and robustness of these methods in more complex environments and with limited map information needs to be further evaluated.
In this project, we would like to develop a real-time feedback framework for this task in parts: perception and detection for obstacles with onboard sensing, and planning for robotic manipulator collision avoidance in cluttered environments e.g forest. Ideally, the algorithm would be evaluated on different robotic platforms for real-time performance, and real-world validation could be carried out.
[1] Jelavic, E., Farshidian, F., Hutter, M., 2021. Combined Sampling and Optimization Based Planning for Legged-Wheeled Robots.
[2] Zhang, X., Liniger, A., & Borrelli, F. Optimization-based collision avoidance.
Obstacle avoidance is a fundamental problem in robotics and is particularly relevant for field robots operating in cluttered environments, e.g. forestry robots. While sampling- and optimization-based methods are widely used for current robotic applications, they are not always able to handle both kinodynamic considerations and arbitrary-shaped obstacles in real-time. Previous research at RSL [1] combined the advantage of both classes of algorithms to achieve efficient and feasible motion planning. Recent advances in MPC algorithms also [2] showed the possibility to deal with nondifferentiable polytopic constraints using dualization. However, the efficiency and robustness of these methods in more complex environments and with limited map information needs to be further evaluated.
In this project, we would like to develop a real-time feedback framework for this task in parts: perception and detection for obstacles with onboard sensing, and planning for robotic manipulator collision avoidance in cluttered environments e.g forest. Ideally, the algorithm would be evaluated on different robotic platforms for real-time performance, and real-world validation could be carried out.
[1] Jelavic, E., Farshidian, F., Hutter, M., 2021. Combined Sampling and Optimization Based Planning for Legged-Wheeled Robots.
[2] Zhang, X., Liniger, A., & Borrelli, F. Optimization-based collision avoidance.
Literature review on planning algorithms
Define the objectives and targets
Implement the planning algorithm and benchmark performance of existing methods
Optional: real-world evaluation tests
Literature review on planning algorithms
Define the objectives and targets
Implement the planning algorithm and benchmark performance of existing methods
Optional: real-world evaluation tests
Highly motivated for the topic
Programming experience (C++/Python)
Knowledge in planning, optimization, and robot dynamics is a plus
Highly motivated for the topic
Programming experience (C++/Python)
Knowledge in planning, optimization, and robot dynamics is a plus
Send an email with your CV and transcript to:
Fang Nan: fannan@ethz.ch
Fan Yang: fanyang1@ethz.ch
Julian Nubert: nubertj@ethz.ch
Send an email with your CV and transcript to:
Fang Nan: fannan@ethz.ch Fan Yang: fanyang1@ethz.ch Julian Nubert: nubertj@ethz.ch
