Figure 8From: A computational method for key-performance-indicator-based parameter identification of industrial manipulators Kernel density estimations of the position precision of the tool center point (TCP) between the new approach using key performance indicator and (a) the heuristic approach and (b) the state-of-the-art optimization approach. The Histogram presents the difference of TCP precision of working configuration \(\boldsymbol {q}^{\mathrm{work}}\) from the new approach using KPI and the heuristic/state-of-the-art optimized measurement configurations \(\boldsymbol {q}_{x}^{\mathrm{meas}}\), where \(x \in\{{h}, \boldsymbol {p}\}\) for 90 parameter identifications differing in the industrial manipulators’ models. Higher values on the x-axis mean a better result of the new approach. The optimized measurement configurations \(\boldsymbol {q}_{p}\) and \(\boldsymbol {q}_{s}\) result by solving the optimization problems: \(\boldsymbol {q}_{p}^{\mathrm{meas}} :=\min_{ \boldsymbol {q}^{\mathrm{meas}}} \operatorname {tr}( {C}_{p}( \boldsymbol {q}^{\mathrm{total}}))\) and \(\boldsymbol {q}_{s}^{\mathrm{meas}} :=\min_{ \boldsymbol {q}^{\mathrm{meas}}} \operatorname {tr}( {C}_{s}( \boldsymbol {q}^{\mathrm{total}}))\).Back to article page