The ROBCOM project aims at the development of an integrated, on-line and off-line, condition monitoring and performance evaluation expert system for industrial robots. The system will be able to detect and classify incipient failures or decalibration trends of the robot, such that production time loss due to curative maintenance can be minimised.
A specific set of robot errors will be studied throughout the project. These errors are related to the following robot subsystems: sensors, arm mechanism, control unit, motor, gear train and brakes. The following parameters are studied: robot endpoint position (ISO 9283 Robot Performance Test Standard), vibration, temperature, forces and electric signals.
Using modern digital signal processing techniques, like singular value decomposition, spectral distance, neural net modelling and fuzzy logic, a simulation of the dynamic model of the robot will be executed for different robot configurations, to study the influence of errors on the robot performance.
Eventually, the ROBCOM-system will be developed into an expert system, providing advice on the type(s) of incipient failures developing in the robot and on the corrective actions to be taken. The result of this project will therefore be a fully integrated error diagnosis system for robot systems and other machinery working in transitory conditions, useful for assessing new equipment as well as for failure detection and predictive maintenance planning during operation.
Funding SchemeCSC - Cost-sharing contracts
2628 BL Delft