为解决目前喷涂作业过程中环境精确感知难、喷涂效果评价难、喷涂路径规划复杂等问题，针对感知识别与路径规划技术进行研究。由于传统喷浆设备喷涂误差累计及不确定性，造成每环喷涂质量参差不齐及回弹量、剥落量居高不下，采用基于轴线归一法的隧道模型构建和多工序扫描策略，实现隧道环境识别、车体定位和喷涂质量感知，提高喷浆机器人外部环境感知能力；
针对传统喷浆设备对多变复杂异构隧道适应能力差，导致无法灵活调整喷涂轨迹问题，提出多工况感知识别及路径规划策略，研究设计最优喷涂路径规划、减少直喷拱架的连续喷涂路径规划方法，降低喷浆机器人与隧道开挖工法的耦合性。在郭达山隧道、任家沟隧道等项目应用，感知识别断面拱架准确率达93%，定位精度、臂姿解算补偿误差控制在50 mm以内，验收质量合格情况下综合回弹率降低至11.9%。
Perception
recognition and path planning technologies are investigated in this study to
address the various issues faced during the spraying process, such as
challenging accurate environmental perception, hard evaluation of wet spraying
effect, and complex spraying path planning. Traditional spraying equipments
have accumulated several unreliable spraying faults, which lead to inconsistent
spraying quality of each ring as well as considerable rebound spalling. To
increase the shotcrete robot′s ability to perceive its external
environment, a tunnel model based on an axis normalization approach and multiprocess scanning strategy is employed to realize the tunnel
environment recognition, vehicle body positioning, and spraying quality
perception. Due to the poor adaptability of traditional slurry spraying
equipment in complex heterogeneous tunnels, the spraying trajectory cannot be
adjusted flexibly. Therefore, a proposal for a multicycle
perception recognition and path planning technique is proposed. Furthermore, to
reduce the coupling between the shotcrete robot and tunnel excavation method,
optimal spraying path planning and continuous spray path planning method are
explored and designed to reduce direct spray arch. Technical achievements have
been used in projects like Guodashan and Renjiagou tunnels, and notable
results, such as a 93% accuracy rate of perception and identification of crosssection arch, a positioning accuracy and arm compensation error
within 50 mm, and a 11.9 % comprehensive rebound rate have been achieved.
