LQG-MP: Optimized path planning for robots with motion uncertainty and imperfect state information

By Jur van den Berg, Pieter Abbeel, Ken Goldberg. 2010

480 citations recorded by the source corpus.

In this paper we present LQG-MP (linear-quadratic Gaussian motion planning), a new approach to robot motion planning that takes into account the sensors and the controller that will be used during the execution of the robot’s path. LQG-MP is based on the linear-quadratic controller with Gaussian models of uncertainty, and explicitly characterizes in advance (i.e. before execution) the a priori probability distributions of the state of the robot along its path. These distributions can be used to assess the quality of the path, for instance by computing the probability of avoiding collisions. Many methods can be used to generate the required ensemble of candidate paths from which the best path is selected; in this paper we report results using rapidly exploring random trees (RRT). We study the performance of LQG-MP with simulation experiments in three scenarios: (A) a kinodynamic car-like robot, (B) multi-robot planning with differential-drive robots, and (C) a 6-DOF serial manipulator. We also present a method that applies Kalman smoothing to make paths Ck-continuous and apply LQG-MP to precomputed roadmaps using a variant of Dijkstra’s algorithm to efficiently find high-quality paths.

Subfields: Automotive Engineering, Computational Theory and Mathematics, Computer Vision and Pattern Recognition, Reinforcement Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Automotive Engineering

In Computational Theory and Mathematics

In Computer Vision and Pattern Recognition

In Reinforcement Learning

Citations in the corpus

Learning Trajectory Preferences for Manipulators via Iterative Improvement (2013, methodology): “[4], we sample trajectories using rapidly-exploring random tree (RRT) [20].”
Learning preferences for manipulation tasks from online coactive feedback (2015, methodology): “More precisely, similar to Berg et al. [6], we sample trajectories using rapidly-exploring random trees (RRT) [33].”
VALUATION : A N ADVERSARIAL APPROACH TO UNCOVER CATASTROPHIC FAILURES (2018, unknown): “In robotics, the assessment of the safety of motion control algorithms has been recognized as a critical aspect of their real-world deployment (Van Den Berg et al., 2011).”
Belief space planning assuming maximum likelihood observations (2010, background): “Van der Berget. al. propose a related approach where a set of potential trajectories are evaluated by tracking the belief state along each trajectory and selecting the one that minimizes th likelihood of failure [9].”
Push-manipulation of complex passive mobile objects using experimentally acquired motion models (2014, background): “Taking uncertainty into account during RRT planning has been studied in the literature (Melchior and Simmons 2007; Berg et al. 2010), however, we do it for the manipulated objects instead of the robot itself in addition to performing it in a novel way.”
Model Predictive Path Integral Control using Covariance Variable Importance Sampling (2015, background): “namics to generate a new nominal state trajectory. This process continues until convergence. There have been several applications of DDP to autonomous systems that demonstrate its broad applicability [9, 10, 11, 12]. There are some notable differences between DDP and iterative path integral control.”
Intention-aware online POMDP planning for autonomous driving in a crowd (2015, unknown): “, [9], [21], [22]).”
Rapidly-exploring Random Belief Trees for motion planning under uncertainty (2011, methodology): “[14] and also derived independently by the authors [15], while He et al.”
Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation (2015, background): “…avoidance models occurs in other fields as well: work on multi-robot coordination in van den Berg et al. (2008), van den Berg et al. (2009), van den Berg et al. (2011) and Snape et al. (2011) shows that robots programmed to jointly avoid each other are guaranteed to be collision free and display…”
Image Object Label 3 D CAD Model Candidate Grasps Google Object Recognition Engine Google Cloud Storage Select Feasible Grasp with Highest Success Probability Pose EstimationCamera Robots Cloud 3 D Sensor (2014, methodology): “For example, parallel Cloud Computing can be used to compute the outcomes of the cross-product of many possible perturbations in object and environment pose, shape, and robot response to sensors and commands [153].”
Integrated task and motion planning in belief space (2013, background): “Under some restricted assumptions, suitable for mobile-robot navigation and control, much more efficient solutions are possible (Prentice and Roy, 2007; van den Berg et al., 2011).”