Robot Arm External Wrench Estimation Without Force Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating external forces and torques on industrial robot arms require additional sensors, leading to high costs and complexity, and existing solutions suffer from low-quality estimates due to manual tuning and reliance on constant friction models.
Innovation Solution
A method that estimates external wrenches at the tool center point using solely joint angles and motor torques, employing a dynamic model and online adaptation of friction models and weighting matrices to improve accuracy and avoid the need for external sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If additional sensors (joint torque sensors or force/torque sensors) are mounted on the robot, then measurement precision of external forces and torques is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses an observer as an intermediary computational element that processes readily available motor torque measurements and robot dynamic model data to estimate external forces and torques. This observer-based approach acts as a mathematical mediator that transforms accessible measurements into the desired force/torque information without requiring direct physical sensing of these quantities.
Solution Approach 2:
The patent replaces the mechanical sensing system (physical force/torque sensors) with a computational estimation system based on the robot's dynamic model and motor torque measurements. This substitution uses mathematical modeling and signal processing to achieve force estimation without the need for complex mechanical sensor installations.
2Measurement precision
If joint torque sensors are used for force estimation, then measurement precision is improved, but ease of manufacture and installation deteriorate
Solution Approach 1:
The robot controller performs force estimation using its own existing resources: the robot dynamic model already stored in memory and motor torque measurements already being collected for control purposes. The system serves its own force estimation needs without requiring external sensing hardware, making the robot essentially self-sufficient for this measurement task.
Solution Approach 2:
The patent makes the robot controller multi-functional by enabling it to perform both motion control and force estimation using the same hardware resources. The robot dynamic model and motor torque measurements serve dual purposes: controlling robot motion and estimating external forces, eliminating the need for dedicated force sensing hardware.
3Ease of operation
If constant weighting matrices are used in force estimation, then ease of operation is improved, but measurement precision deteriorates due to manual tuning requirements
Solution Approach 1:
The patent transforms the static, constant weighting matrices into dynamic, time-varying matrices that are continuously adapted based on the current robot state and operating conditions. This dynamic adaptation allows the estimation algorithm to optimize its performance for different scenarios automatically, improving measurement precision without requiring manual retuning.
Solution Approach 2:
The patent implements feedback mechanisms where the observer continuously monitors the robot's actual behavior and compares it with model predictions, using this information to adaptively adjust the weighting matrices. This feedback-driven adaptation enables the system to maintain high estimation accuracy across varying operating conditions without manual intervention.
Data Source
Figure 1
AI summary
The invention is related to a method for estimating the externally applied wrench (12, fext) on the robot arm (46) caused by collision, based on the calculation of the sum of all the torques applied (e.g. joint motor, friction, gravity, dynamic behaviour) and wherein the robot comprises a robot controller (28) which is foreseen to control the movement of the robot arm (46) according to data of a robot program and to continuously provide: the torque vector of the joint motors (τmot), the dynamic behavior of the robot arm, a torque vector due to gravity effects (τgrav(q)) and the current joint angles (q) of the robot arm. The method is characterized by the following steps: moving the robot arm (46) according to the data of a robot program, continuously determining the friction torque vector (τfric), continuously determining the external torque vector (τext) by use of the following equation: (A) providing the Jacobian Matrix (J) of the robot arm and continuously estimating the externally applied wrench (fext) of the robot arm (46) dependent on the angle vector (q) based on the Jacobian matrix (J) and the external torque vector (τ ext ).