Tactile End-Effector Control for Robotic Manipulation Under Occlusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional vision-based robotic manipulation systems face limitations in dexterous tasks due to susceptibility to occlusion and other challenges, such as poor lighting, making them less effective in environments where detailed and varied sensor feedback is needed.

Innovation Solution

A tactile sensing-based robotic system that uses end-effectors with tactile sensors to sense contact with objects, control contact states, estimate object poses, and plan trajectories based on tactile data, employing manipulation primitives like grasp, push, pivot, and pull to achieve robust control and adapt to external perturbations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If vision-based feedback control is used for robotic manipulation, then the system can perform basic pick-and-place tasks, but it fails to achieve dexterous manipulation due to occlusion and lighting limitations

Engineering Contradiction:
Improvedexterous manipulation capabilityVSAvoidperformance under occlusion and poor lighting
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent replaces vision-based sensing with tactile sensing at the end effector. Tactile sensors directly measure contact forces, positions, and object properties through mechanical interaction, eliminating dependence on visual conditions. This substitution enables reliable dexterous manipulation by sensing through physical contact rather than optical fields that are blocked by occlusions or affected by lighting.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces tactile sensors as an intermediary between the end effector and the object. These sensors act as a mediator that translates physical contact into actionable feedback signals, enabling the control system to perceive contact states, estimate object poses, and adjust manipulation strategies in real-time without relying on vision systems.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If tactile sensors are deployed at the end effector for detailed contact sensing, then dexterous manipulation and adaptability improve, but system complexity increases

Engineering Contradiction:
Improvetactile feedback capabilityVSAvoidend effector sensor integration
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent designs the tactile sensor system to perform multiple functions simultaneously: detecting contact presence, measuring contact forces, estimating object pose, and providing feedback for trajectory planning. By consolidating these sensing capabilities into a single integrated tactile sensor array at the end effector, the system achieves multi-functionality that reduces overall system complexity compared to using separate sensors for each function.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The tactile sensors enable the robotic system to self-regulate its manipulation actions based on real-time contact feedback. The system automatically adjusts contact forces, modifies trajectories, and adapts to object properties without external intervention, making the complex sensor integration serve the system's own control needs autonomously.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS11878415B2Tactile dexterity and control
Publication Date: 2024.01.23 MASSACHUSETTS INST OF TECH
  • US11878415B2 patent drawing
  • US11878415B2 patent drawing
  • US11878415B2 patent drawing

AI summary

Systems and methods relating to tactile dexterity and control are disclosed. In one embodiment, a method of manipulating an object based on tactile sensing includes sensing an object by receiving signals from a tactile sensor of an end effector of a robotic system in contact with the object, controlling a contact state by operating the end effector to enforce a desired contact condition between the end effector and the object, estimating a pose of the object based on the received signals, and planning at least one trajectory of the object based on the estimated pose of the object and a desired pose of the object.