Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Robot's end effector inertial forces compensation for better collision detection
Using force/torque and acceleration data of robot end effector to isolate external forces by canceling out tool's Inertial forces. This compensation will enhance collision detection to make robot manipulators safe to perform collaborative tasks.
As collaborative robots are robots designed to operate side by side with humans a large area of interest is safety. Making a robot safe often means reducing performance to be able to brake in time or minimize collision forces. Many robots these day are equipped with force/torque sensors to enhance their dexterity in tasks like assembly, polishing/grinding, surface detection. These sensors despite their very good sensitivity and accuracy, often suffer from drift (caused by temperature or loading of the body) in the measurements making them unusable to measure absolute loads and requires to be reference quite often
The goal of this project is to use the force/torque measurements and combine them with acceleration data to compensate for inertial forces when performing dynamic movements. This will allow the robot to detect much smaller external forces during a task so making safer for interaction. Also, the other way, when idling or in less dynamic movements to use this data to continuously evaluate and correct the force/torque measurement's drift.
This will lead to a force/torque measuring device that can be actually be considered almost ideal from the higher level control perspective, allowing less complex control algorithms and easier integration of such devices in robotics systems.
You will:
- Have the opportunity to work with different types of robots including ANYmal robot
- familiarize yourself with the control of modern complex robots
- familiarize with collaborative robotics and their potential.
- be part of a big and very active Lab in the robotics field.
As collaborative robots are robots designed to operate side by side with humans a large area of interest is safety. Making a robot safe often means reducing performance to be able to brake in time or minimize collision forces. Many robots these day are equipped with force/torque sensors to enhance their dexterity in tasks like assembly, polishing/grinding, surface detection. These sensors despite their very good sensitivity and accuracy, often suffer from drift (caused by temperature or loading of the body) in the measurements making them unusable to measure absolute loads and requires to be reference quite often
The goal of this project is to use the force/torque measurements and combine them with acceleration data to compensate for inertial forces when performing dynamic movements. This will allow the robot to detect much smaller external forces during a task so making safer for interaction. Also, the other way, when idling or in less dynamic movements to use this data to continuously evaluate and correct the force/torque measurement's drift.
This will lead to a force/torque measuring device that can be actually be considered almost ideal from the higher level control perspective, allowing less complex control algorithms and easier integration of such devices in robotics systems.
You will:
- Have the opportunity to work with different types of robots including ANYmal robot - familiarize yourself with the control of modern complex robots - familiarize with collaborative robotics and their potential. - be part of a big and very active Lab in the robotics field.
1. Design and implement the algorithm in any programming language.
2. Evaluate the algorithm using logged data from various robots and define the parameters that affect the quality of the outcome.
3. Implement the algorithm in low level inside the force/torque sensor device software that will be specifically designed by us to optimize the algorithm's outcome
1. Design and implement the algorithm in any programming language. 2. Evaluate the algorithm using logged data from various robots and define the parameters that affect the quality of the outcome. 3. Implement the algorithm in low level inside the force/torque sensor device software that will be specifically designed by us to optimize the algorithm's outcome
- experience in programming in C/C++ or Python
- experience with Robot operating system
- experience in robot dynamics and kinematics
- experience in programming in C/C++ or Python - experience with Robot operating system - experience in robot dynamics and kinematics
Please send your CV to Ilias Patsiaouras:
ilias.patsiaouras@mavt.ethz.ch