Design of and Experimentation on an Intelligent Intra-Row Obstacle Avoidance and Weeding Machine for Orchards

Based on the current issues of difficulty in clearing intra-row weeds in orchards, inaccurate sensor detection, and the inability to adjust the row spacing depth, this study designs an intelligent intra-row obstacle avoidance and weeding machine for orchards. We designed the weeding machine’s sensor device, depth-limiting device, row spacing adjustment mechanism, joystick-based obstacle avoidance mechanism, weeding shovel, and hydraulic system. The sensor device integrates non-contact sensors and a mechanical tactile structure, which overcomes the instability of non-contact detection and avoids the risk of collision obstacle avoidance by the weeding parts. The weeding shovel can be adapted to the environments of orchards with small plant spacing. The combination of the sensor device and the obstacle avoidance mechanism realizes flexible obstacle avoidance. We used Ansys Workbench to conduct static and vibration modal analyses on the chassis of the in-field weeding machine. On this basis, through topology optimization, the chassis quality of the weeding machine is reduced by 8%, which realizes the goal of light weight and ensures the stable operation of the machinery. To further optimize the weeding operation parameters, we employed the Box–Behnken design response surface analysis, with weeding coverage as the optimization target. We systematically explored the effects of forward speed, hydraulic cylinder extension speed, and retraction speed on the weeding efficiency. The optimal operational parameter combination determined by this study for the weeding machine is as follows: forward speed of 0.5 m/s, hydraulic cylinder extension speed of 11.5 cm/s, and hydraulic cylinder retraction speed of 8 cm/s. Based on the theoretical analysis and scenario simulations, we validated the performance of the weeding machine through field experiments. The results show that the weeding machine, while exhibiting excellent obstacle avoidance performance, can achieve a maximum weeding coverage of 84.6%. This study provides a theoretical foundation and technical support for the design and development of in-field mechanical weeding, which is of great significance for achieving intelligent orchard management and further improving fruit yield and quality.