Motion Planning Engineer, Autonomous Driving
At Ridecell View All Jobs
Motion Planning Engineer
Motion planning is a core function in an autonomous driving system. Given a destination we want to reach, can you figure out how to make your way through a dynamic road environment? Taking into account all the constraints we must meet - following traffic regulations, avoiding other vehicles and pedestrians on the road, and accounting for the uncertainty associated with their future behavior.
Join us today to work on this fascinating problem, and design innovative solutions that produce optimal yet safe driving behavior!
What You'll Do
- Develop lane-based planners for structured environments such as public roads.
- Develop open-space planners for navigating in unstructured environments.
- Incorporate map data, including road boundaries, traffic elements, speed limits.
- Model and predict obstacle motion, estimate and account for uncertainty.
- Generate high-level behaviors and actions using state machines, behavior trees.
- Generate, optimize and validate obstacle-free trajectories.
- Design hybrid planning architectures with rule-based and learned agents.
- Interface with higher-level routing and decision-making modules.
- Interface with lower-level control and actuation modules.
- Bachelors or Masters degree in Computer Science, Robotics, Electrical, Electronics, Mechanical or Aerospace Engineering, or a related field.
- 2+ years experience developing motion planning modules for autonomous vehicles, aerospace or other safety-critical robotic systems.
- Proficiency in trajectory generation, optimization and validation techniques.
- Proficiency in applying linear algebra, probability theory and numerical optimization.
- Proficiency in developing object-oriented code and testing routines in C++.
- Experience with search-based algorithms such as A* and RRT* with motion primitives.
- Experience with polynomial and spline fitting methods.
- Experience with robust, time-critical and compute-constrained implementation techniques.
Nice to Have
- Experience working with robotics middleware (e.g. ROS) and simulation platforms.
- Familiarity with vehicle dynamics, Model-Predictive Control (MPC).
- Familiarity with POMDPs, Dynamic Programming and Reinforcement Learning.
You do not need to match every desired qualification to apply. This is a very interdisciplinary field of engineering, and we enjoy working with passionate engineers coming with diverse, overlapping, and unique skill sets.