Accurate Indirect 3-D Localization of Markerless Industrial Products

This paper proposes a novel method for the indirect, automatic, real-time, and three-dimensional tracking of industrial products. The motivation behind the proposed method is to fully avoid the direct product marking with active or passive tags. This method combines the tracked forklift position and heading data with the real-time fork elevation to obtain a complete 3D location of the transported product. This work also proposes a fusion algorithm for automatic product pick-up and drop-down event detection (A-PDD). It is based on the fusion of the fork elevation and occupancy sensors, used for the automatic recognition of the forklift behavior patterns during the payload pick-up or drop-down events. This work experimentally validates the performance of the proposed markerless indirect tracking method in an operating industrial production site. The proposed method was able to indirectly track two industrial products with below 30 cm absolute and repeatability errors. The experimental comparison of the proposed indirect tracking method with the traditional direct product localization approach has shown a major advantage of the proposed method in terms of product positioning stability, reliability, and quality, along with approximately four times higher product positioning accuracy. The proposed indirect tracking method offers high scalability, significantly increased cost- and energy efficiency, and reduced maintenance requirements in comparison with the direct tracking methods. It is expected to be widely applicable to different industrial machinery and material handling equipment, from forklifts and various lifters to cranes.