A Novel Approach for Multi-Criteria Workspace Optimization of Parallel Kinematic Machines

Parallel kinematic machines (PKM) or parallel robots have been a topic of research for the last two decades. Serial robots, due to their inherent drawbacks (Higher error amplification at the end effector, Greater weight to payload ratio, etc.), cannot be used in applications that require high accuracy across the workspace. Parallel robots overcome these drawbacks, making them a natural replacement for serial robots. However, they also have associated drawbacks (Small irregularly shaped workspace, many singularities in the workspace, etc.). For complex applications, the use of parallel robots instead of serial robots is preferred. For large-scale deployment of parallel manipulators in industries, remedial measures to overcome their drawbacks need to be developed. Most importantly, workspace optimization is required to exploit the full potential of these robots. This optimization is subject to one or more performance parameters (objectives) that often have conflicting requirements i.e. improving one objective deteriorates the performance of another, which is by far a severe challenge in this domain. An example of conflicting goals is “To purchase a car that is cheap, comfortable, safe, environment-friendly, high power and fuel-efficient”. In such cases, “best compromise between conflicting goals” is the best solution. Thus far, workspace optimization of parallel robots focuses on optimization concerning a single objective at a given time. The basic idea of this work is to propose a methodology to optimize the workspace of a PKM subject to multiple objectives simultaneously. Several optimization schemes including weighted averages, gradient descent, surrogate optimization, and genetic algorithms were studied. Pareto front optimization appears to be most suited to the application at hand.