In robotics research, it is often difficult to compare and evaluate techniques experimentally. This paper identifies these difficulties and provides solutions based on our work during the last four years in the field of sampling-based motion planning.
Roland Geraerts. On Experimental Research in Sampling-based Motion Planning. In (IROS'06) Workshop on Benchmarks in Robotics Research , pp. 31-34, 2006.
The (Windows) framework can be downloaded on our software page.