Robot Gripper Slip Detection for Adaptive Clamping Force
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic grippers face challenges in adapting clamping forces to prevent slippage without causing damage or excessive energy consumption, especially with multi-finger grippers, and existing methods are complex and costly.
Innovation Solution
A robotic gripper with sensors that detect relative movement and a control system using machine learning to adjust clamping forces based on slip detection scores, iteratively and incrementally, without requiring force or friction measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the clamping force is increased to prevent slippage, then the grip stability is improved, but the energy consumption increases and the object may be damaged
Solution Approach 1:
The patent implements a feedback control system that uses sensors to detect slippage and dynamically adjusts the clamping force. The control system receives sensor data indicating relative movement between the gripper and object, processes this information through machine learning algorithms to determine slip detection scores, and adjusts the clamping force accordingly. This feedback mechanism ensures the gripper applies only the necessary force to maintain grip stability, avoiding excessive energy consumption and object damage.
Solution Approach 2:
The patent employs dynamic adjustment of clamping force based on real-time slip detection. Instead of applying a constant high clamping force, the system dynamically modulates the force level according to detected slippage conditions. The machine learning model iteratively adjusts the clamping force in response to changing grip conditions, allowing the gripper to adapt to varying object properties and external disturbances while minimizing energy consumption.
2Reliability
If the clamping force is increased to prevent slippage, then the grip stability is improved, but the object may be damaged by excessive force
Solution Approach 1:
The feedback control system continuously monitors grip conditions through sensors and adjusts clamping force in real-time. By detecting slippage and responding with minimal necessary force increases, the system maintains grip stability while avoiding excessive forces that could damage the object. The machine learning algorithm processes sensor data to determine appropriate force adjustments, ensuring forces remain within safe limits.
Solution Approach 2:
The patent changes the clamping force parameter dynamically based on detected slippage conditions. Rather than using a fixed high force, the system adjusts the force parameter in response to slip detection scores generated by the machine learning model. This parameter modulation ensures sufficient grip stability while keeping forces low enough to prevent object damage.
3Reliability
If traditional force optimization methods are used to determine minimum forces, then the grip stability is improved, but the device complexity and cost increase
Solution Approach 1:
The patent replaces complex mechanical force measurement and optimization systems with a sensor-based detection and machine learning approach. Instead of using multiple force sensors and complex GRASP matrix calculations, the system uses simpler sensors to detect relative movement and employs machine learning algorithms to determine appropriate clamping forces. This substitution reduces device complexity and cost while maintaining grip stability.
Solution Approach 2:
The patent uses machine learning models trained on simulation data to replicate the functionality of complex force optimization methods. The neural network is trained offline using GRASP matrix calculations and force optimization algorithms, then deployed as a lightweight model that can quickly determine appropriate clamping forces without performing complex real-time calculations. This copying approach transfers the intelligence of complex methods to a simpler implementation.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A robotic gripper comprising: One or more fingers, at least one actuator for acting on the fingers to grip an object, the gripper comprising at least one sensor sensitive to a relative movement between the object manipulated by the gripper and the latter, this sensor delivering an output signal, - a control system configured to execute at least one machine learning algorithm trained to deliver from the output signal(s) from the sensor(s) at least one slip detection score representative of the algorithm's confidence in the presence of slippage, and to transmit control data to the actuators controlling the finger(s), the control data being generated so as to link the intensity of the forces applied by the finger(s) to said score.