The working environment of a picking robot is complex, and the motion-planning algorithm of the picking manipulator will directly affect the obstacle avoidance effect and picking efficiency of the manipulator. In this study, a time-optimal rapidly-exploring random tree (TO-RRT) algorithm is proposed. First, this algorithm controls the target offset probability of the random tree through the potential field and introduces a node-first search strategy to make the random tree quickly escape from the repulsive potential field. Second, an attractive step size and a “step-size dichotomy” are proposed to improve the directional search ability of the random tree outside the repulsive potential field and solve the problem of an excessively large step size in extreme cases. Finally, a regression superposition algorithm is used to enhance the ability of the random tree to explore unknown space in the repulsive potential field. In this paper, independent experiments were carried out in MATLAB, MoveIt!, and real environments. The path-planning speed was increased by 99.73%, the path length was decreased by 17.88%, and the number of collision detections was reduced by 99.08%. The TO-RRT algorithm can be used to provide key technical support for the subsequent design of picking robots.

The soluble solid content (SSC) is one of the important parameters depicting the quality, maturity and taste of fruits. This study explored hyperspectral imaging (HSI) and fluorescence spectral imaging (FSI) techniques, as well as suitable chemometric techniques to predict the SSC in kiwifruit. 90 kiwifruit samples were divided into 70 calibration sets and 20 prediction sets. The hyperspectral images of samples in the spectral range of 387 nm~1034 nm and the fluorescence spectral images in the spectral range of 400 nm~1000 nm were collected, and their regions of interest were extracted. Six spectral pre-processing techniques were used to pre-process the two spectral data, and the best pre-processing method was selected after comparing it with the predicted results. Then, five primary and three secondary feature extraction algorithms were used to extract feature variables from the pre-processed spectral data. Subsequently, three regression prediction models, i.e., the extreme learning machines (ELM), the partial least squares regression (PLSR) and the particle swarm optimization - least square support vector machine (PSO-LSSVM), were established. The prediction results were analyzed and compared further. MASS-Boss-ELM, based on fluorescence spectral imaging technique, exhibited the best prediction performance for the kiwifruit SSC, with the Rp2, Rc2 and RPD of 0.8894, 0.9429 and 2.88, respectively. MASS-Boss-PLSR based on the hyperspectral imaging technique showed a slightly lower prediction performance, with the Rp2, Rc2, and RPD of 0.8717, 0.8747, and 2.89, respectively. The outcome presents that the two spectral imaging techniques are suitable for the non-destructive prediction of fruit quality. Among them, the FSI technology illustrates better prediction, providing technical support for the non-destructive detection of intrinsic fruit quality.

This paper presents the results of a motion planning algorithm that has been used in an intelligent citrus-picking robot consisting of a six-link manipulator. The real-time performance of a motion planning algorithm is urgently required by the picking robot. Within the artificial potential field (APF) method, the motion planning of the picking manipulator was basically solved. However, the real-time requirement of the picking robot had not been totally satisfied by APF because of some native defects, such as the large number of calculations used to map forces into torques by the Jacobian matrix, local minimum trap, and target not reachable problem, which greatly reduce motion planning efficiency and real-time performance of citrus-picking robots. To circumvent those problems, this paper proposed some novel methods that improved the mathematical models of APF and directly calculates the attractive torques in the joint space. By using the latter approach, the calculation time and the total joint error were separately reduced by 54.89% and 45.41% compared with APF. Finally, the novel algorithm is presented and demonstrated with some illustrative examples of the citrus picking robot, both offline during the design phase as well as online during a realistic picking test.

No abstract

No abstract