EASYRob - Efficient and advanced system and methods for the elasto-kinematic calibration of robots

Background

Increasing product complexity and variety leads to a strong demand for more flexible production technology. However, reusing the already existing technologies and materials is the fundamental building block to sus-taining the quality of our western life and at the same time facilitates the movement towards sustainable soci-eties. We already see a strong trend in using digital processes to optimize both the design and commissioning of nearly all types of production systems. Of very high interest are industrial manipulators (robots), as they are low-cost, resource efficient designs with highest flexibility and capability. The downside with robots is their insufficient accuracy in the context of manufacturing. This comes with downsides like:

• Limited possibilities to offline programming and true virtual commissioning. Installations need to be manually tuned or adapted on-site by skilled staff.
• Additional equipment is added to lower the accuracy requirements, adding cost and complexity to the end solution.
• Poor predictability of the robot behavior adds risk to the installation.
• Lack of ability to reach required application tolerances and to know which tolerances can be reached.

The result is an unnecessarily high cost for manual tuning, which often kills the business case in high salary countries. This results in manual production placed in low salary countries, or that less flexible machines are chosen over general purpose robots, leading to more resource consumption and higher costs when needing to adapt to product changes.

Calibration is the solution to increase a robot’s absolute accuracy, and elasto-kinematic calibration makes it applicable to value adding applications and takes the performance prediction possibilities to the next level.

Absolute accuracy options have been available from most robot manufacturers for decades, but they are only chosen for a very small part (<5%) of the robot population. The reason being the price, unclear resulting accuracy, lack of understanding for the provided value, and lack of possibility to recover in situations of failure. Being an option when ordering the robot, it is many times too late when the user realizes the need. The variety of robot cells and the lack of space within a cell make it many times too costly to add the func-tionality since the robot must be taken out of the production environment for a costly calibration made by expert service providers.

The goal of this project is to create a fundamentally new way of providing robot calibration to the market.

This is primarily achieved via an extremely robust and portable measurement device, packaged in a smart way for the applicability by all robot users everywhere. The second part is a software option that assists to carry out measurements within the limited space in the robot cell and that offers transparency in the reachable meas-urement uncertainty. Field calibration (calibration at robot site) and thereby accuracy on demand is thereby made attractive. The system will be usable on both new and already installed robots. With this, robots can take a leading role in value adding applications, e.g., making local small-scale production profitable and thereby contributing to the sustainable future we all need.

Project description

With a consortium of four experienced partners in metrology systems and robot calibration, we want to create a new measurement concept that can be added on demand by any robust user or integrator. Apart from being robust, the concept includes the generation of load cases to directly deliver data for the identification of the robot component deflections (elasto-kinematic calibration) without the need for additional components. The latter is the key for revolutionizing accuracy by keeping the robot accurate also for heavy tooling, varying load cases and by giving clarity on the resulting process accuracy. The approach is accompanied by software solutions to ease the field application by providing dialogue-based measurement planning, advanced visuali-zation tools, and safety measures such as collision avoidance and EC compliance. The described approach makes it superior to the existing market solution.

Aim and objectives

The project goal is to make elasto-kinematic calibration accessible to all robot us-ers, enabling them to repurpose existing equipment for demanding applications and manufacturing processes beyond low-force and low-accuracy tasks. This is achieved by combining the experience of the consortium in the fields of large-scale robot and force measurements (IWU, KTH) as well as precision design (OPW) and robot calibration (CB).
Over the course of this project, the following topics are researched:

• Fraunhofer IWU
  • Real time measurement of device and robot deflections under force loads
  • Generation of optimal, collision free trajectories with maximized calibration sensitivity
• OPW GmbH
  • Realization of high accuracy measurements of forces and displacements across large scales
  • Design and production of robust and field-applicable measurement equipment
• KTH
  • Influence of the feedback between robot and measurement device on the measurement uncertainty
  • Prediction of the system capability for guiding the measurement planning
• Cognibotics AB
  • Optimization of elasto-kinematic robot models and parameter identification algorithms
  • Packaging of hard- and software solutions for easy applicability to the large robot user base

Outcomes

The main outcomes of the project are new methods and harware for the elasto-kinematic calibration of industrial robots. These tools will be applicable on-site and are accompanied by software allowing simple robot calibration for a wide range of the installed industrial robot base.

Funded by

Vinnova (Sweden) and ZIM (Germany)

Duration

August 2023 – August 2025

Project partners

• KTH Royal Institute of Technology
• Fraunhofer IWU
• Cognibotics AB
• Oberndorfer Präzisions-Werk GmbH & Co. KG

Contact

Project leader

Projekt partner