LAUSR

An analysis of the existing hand-eye calibration methods reveals that most of them are far from trivial. And, what is worse, their intrinsic complexity makes it difficult to elucidate under which circumstances they fail to provide an accurate solution. Thus, although it might seem that the hand-eye calibration problem is uninspiring because it is assumed to be well-solved, we show in this letter that there was still room for improvement, both in terms of simplicity and robustness. After reviewing the most representative methods, we analyze the situations in which they fail, and we introduce a simpler closed-form alternative that accurately solves the problem in all the identified critical circumstances. Its performance is evaluated using simulated and real experimental data.