The surface of the Moon and planets have been covered with loose soil called regolith, and there is a risk that the rovers may stack, so it is necessary for them to recognize the traveling state such as its posture, slip behavior, and sinkage. There are several methods for recognizing the traveling state such as a system using cameras and Lidar, and they are used in real exploration missions like Mars Exploration Rovers of NASA/JPL. When a rover travels and travels across loose soil with steep slopes like a side wall of a crater on the lunar surface, the rover has side slipping. It means that its behavior makes the rover slip down to the valley direction. Even if this detection uses sensors like a camera and Lidar or other controlling systems like SLAM (Simultaneous Localization and Mapping), it would be too difficult for the rover to avoid slipping down to valley direction, because it is not able to detect the traction or resistance given from ground by individual wheel of the rover, as the traction of individual wheel of the rover is not clear. This means that the movement of the rover appeared by integrating the traction of all wheels mounted on the rover. Even if the localization by sensors is carried out, the location would be the location after slipping down. This is because when traveling on unstable ground, the driving force of each individual wheel cannot be accurately predicted, and the sum of the driving force of all wheels is the motion of the rover, which is detected after the position changes. Therefore, if the rover obtains information on the traction of each wheel, its maneuver to change its posture would work sooner and it would be able to travel more efficiently than in a state without that information. Because the onboard computer of rovers can identify their location and state from the information of the traction of each wheel, they can decide the next work carefully and in detail. From these tasks, we focused on the intrinsic sensation of a biological function like a human body and aimed to develop a system that recognizes the traveling state (slip condition) from the shape deformation of the chassis. In this study, we experimentally verified the relationship between the change in strain, which is the amount of deformation acting on the chassis, and the traveling state while the wheel is traveling. From the experimental results, we confirmed that the strain in the chassis was displaced dynamically and that the strain changed oscillatory while the wheel was traveling. In addition, based on the function of muscle spindles as mechanoreceptors, we discussed two methods of analyzing strain change: nuclear chain fiber analysis and nuclear bag fiber analysis. These analyses mean that the raw data of the strain are updated to detect the characteristic strain elements of a chassis while the wheel is traveling through loose soil. Eventually, the slipping state could be estimated by updating the data of a lot of strain raw data, and it was confirmed that the traveling state could be detected.
Read full abstract