UR5 External Force Estimation

• Implement the external force estimation algorithm based on disturbance observer in UR5 in Copeliasim.

Implementation Detail

• Based on the dynamic parameter identification result of the UR5 robot in he Coppeliasim.
• Simulation data containing position, velocity, acceleration and joint torque is collected in the Coppeliasim with the physic engine Bullet2.83. In simulation, there are no friction.
• All data is collected without any external force. Apply the external force artificially in the code, then use the disturbance observer to estimate the external force.

Command

# Dependency to plot the result.
sudo apt-get install gnuplot

# Compile
cd forceEstimation
mkdir build
cmake ..
make

# Usage
./bin/UR5EstimateForceEstimation -h                                           
Allow options:
  -h [ --help ]                        Turn the gain parameter for external 
                                       observer.
  -e [ --externalTorqueType ] arg      external torque type: const or vary
  -p [ --parameterType ] arg (=normal) parameter type: normal, small or big.
  -i [ --inputFile ] arg               input file for external observer.
  -o [ --outputFile ] arg              output file for saving data.
  -f [ --frequency ] arg (=100)        sample Frequency
  -t [ --observerType ] arg            observer type:
                                            0: momentum observer
                                            1: nonlinear observer
                                            2: sliding mode observer
                                            3: filtered dynamic observer
                                            4: kalman filter observer(Tayler)
                                            5: kalman filter observer(Zero 
                                       order filter)

# Example
./bin/UR5EstimateForceEstimation -e vary -p normal -i 6_2000.csv -o 111.csv -t 2

Example

Method	Result
Momentum Observer
Nonlinear Observer
Sliding Mode Observer
Filtered Dynamic Observer
Kalman Filter Observer (Tayler)
Kalman Filter Observer (Zero Order)

Reference

• ext_observer
• Research on Dynamic Parameter Identification and Admittance Control of Collaborative Robots: Chapter3
• Cartesian Contact Force Estimation for Robotic Manipulators using Kalman Filters and the Generalized Momentum
• A Nonlinear Disturbance Observer for Robotic Manipulators
• Motor-Current-Based Estimation of Cartesian Contact Forces and Torques for Robotic Manipulators and Its Application to Force Control
• Discrete-Time Solutions to the Continuous-Time Differential Lyapunov Equation With Applications to Kalman Filtering
• Estimating Robot End-Effector Force from Noisy Actuator Torque Measurements
• Sliding Mode Momentum Observers for Estimation of External Torques and Joint Acceleration