Autonomous pallet localization and picking for industrial forklifts: a robust range and look method

Abstract

A combined double-sensor architecture, laser and camera, and a new algorithm named RLPF are presented as a solution to the problem of identifying and localizing a pallet, the position and angle of which are a priori known with large uncertainty. Solving this task for autonomous robot forklifts is of great value for logistics industry. The state-of-the-art is described to show how our approach overcomes the limitations of using either laser ranging or vision. An extensive experimental campaign and uncertainty analysis are presented. For the docking task, new dynamic nonlinear path planning which takes into account vehicle dynamics is proposed.

Detection, localisation and tracking of pallets using machine learning techniques and 2D range data

“…Other methods for 2D data segmentation, feature detection, fitting, and matching have been presented in [9,52], but all these approaches are characterised by the same drawbacks. An integrated laser and camera sensory system, for solving the problem of simultaneously identifying and localising pallets whose location is characterised by a great uncertainty, has been presented in [7]. However, such an approach suffers from a number of drawbacks typically associated with vision processing.…”

A Method for Autonomous Picking of Paper Reels

“…Therefore, more reliable, cost-effective and suitable sensors must be employed together with robust real-time algorithms. On the other hand, "faulty situations may also occur in highly repetitive and pre-determined tasks when, for example, a pallet lies in an unexpected position and a blind pick-up would cause a system fault and possible damage" [4].…”

“…In reality, the concept of automating logistic must means that all aspect of the completely automated material handling solution, including the loading/unloading phase are addressed. Thus, "industrial autonomous vehicles must be able to understand their surroundings and to cope with uncertainty" [4].…”

“…Research in the field of material transportation logistics and the need for picking algorithms is only one topic in the grasping and picking paradigm. Most research papers which discussed industrial autonomous forklifts were focused on traditional pallet picking [4], [5]- [7]. However, the need for handling materials of different kind and shape such as paper reels in paper mills encouraged researchers [8] to create a modified autonomous forklift by changing the forks into cylindrical clamps.…”

Position-based visual servoing for pallet picking by an articulated-frame-steering hydraulic mobile machine

“…The main challenge in this task is the reliable detection of the pallet while eliminating typical failures associated with vision systems. To increase the robustness of vision, previous studies have mostly used auxiliary sensors such as laser pointers [1] or laser scanners [2]; another study [3] applied a neural extended Kalman filter for object detection. In this study, we propose a fault-tolerant pallet pose (position and orientation) estimation in which vision data is fused with odometry and inertial measurements.…”