Advancements in Robot Motion Planning and Control

The field of robotics is witnessing significant advancements in motion planning and control, with a focus on developing more efficient, adaptive, and robust systems. Researchers are exploring new approaches to represent robot kinematic reachability, such as differentiable reachability maps, to reduce computational costs and improve motion planning. Additionally, there is a growing interest in using model predictive control and optimization techniques to generate real-time whole-body motions for legged robots and humanoid robots. These techniques enable robots to navigate complex environments and perform dexterous manipulation tasks with increased precision and reliability. Furthermore, the development of unified hierarchical control frameworks and three-level whole-body disturbance rejection control frameworks is enhancing the stability and robustness of legged robots in the presence of uncertainties. Noteworthy papers in this area include: Learning Differentiable Reachability Maps for Optimization-based Humanoid Motion Generation, which introduces a new approach to representing robot kinematic reachability. Control of Legged Robots using Model Predictive Optimized Path Integral, which demonstrates a sampling-based model predictive strategy for generating real-time whole-body motions. Geodesic Tracing-Based Kinematic Integration of Rolling and Sliding Contact on Manifold Meshes for Dexterous In-Hand Manipulation, which presents an integration scheme for roll-slide contact modeling on manifold meshes. Simultaneous Contact Sequence and Patch Planning for Dynamic Locomotion, which proposes a full pipeline for simultaneous contact sequence and patch selection. Scaling Whole-body Multi-contact Manipulation with Contact Optimization, which provides a representation of robot and object surfaces for efficient whole-body manipulation planning. Accelerating Signal-Temporal-Logic-Based Task and Motion Planning of Bipedal Navigation using Benders Decomposition, which presents a method for accelerating task and motion planning under Signal Temporal Logic constraints. Unified Hierarchical MPC in Task Executing for Modular Manipulators across Diverse Morphologies, which proposes a unified Hierarchical Model Predictive Control for modular manipulators. A Three-Level Whole-Body Disturbance Rejection Control Framework for Dynamic Motions in Legged Robots, which presents a control framework for enhancing the stability and robustness of legged robots. Task and Motion Planning for Humanoid Loco-manipulation, which presents an optimization-based task and motion planning framework for humanoid loco-manipulation. Hardware Implementation of a Zero-Prior-Knowledge Approach to Lifelong Learning in Kinematic Control of Tendon-Driven Quadrupeds, which demonstrates a bio-inspired learning algorithm for kinematic control of tendon-driven quadrupeds.

Sources

Learning Differentiable Reachability Maps for Optimization-based Humanoid Motion Generation

Control of Legged Robots using Model Predictive Optimized Path Integral

Geodesic Tracing-Based Kinematic Integration of Rolling and Sliding Contact on Manifold Meshes for Dexterous In-Hand Manipulation

Simultaneous Contact Sequence and Patch Planning for Dynamic Locomotion

Scaling Whole-body Multi-contact Manipulation with Contact Optimization

Accelerating Signal-Temporal-Logic-Based Task and Motion Planning of Bipedal Navigation using Benders Decomposition

Unified Hierarchical MPC in Task Executing for Modular Manipulators across Diverse Morphologies

A Three-Level Whole-Body Disturbance Rejection Control Framework for Dynamic Motions in Legged Robots

Task and Motion Planning for Humanoid Loco-manipulation

Hardware Implementation of a Zero-Prior-Knowledge Approach to Lifelong Learning in Kinematic Control of Tendon-Driven Quadrupeds

Built with on top of