A compositive sensing method for autonomous indoor locomotion has been tested and was verified to be effective.
Many sensing data are required if mobile robot is to autonomously climb stairways and vertical steps. Since payload, space and electric power of a mobile robot are strictly limited, sensory hardware must be minimized. If sensors are used in a compositive manner, more information can be obtained than could be from the sensors considered individually, and some of this information cannot be obtained from individual sensors. This compositive sensing method has been applied to the locomotion control of the transformable crawler. First, the appropriate sensors for detecting indoor environments are selected, and sensing methods are discussed. Next, these methods are tested under specified environmental conditions. Lastly, this vehicle was tested by autonomously touring an indoor model passage including stairways and vertical steps. This method was shown to be effective. For this experiment, 6 different kinds and a total of 26 sensors were used, and 27 different kinds and a total of 42 information data were used for locomotion control.