Group control system and group control method
The group control system controls a plurality of mobile objects capable of autonomously traveling in a predetermined area. The group control system includes a position information estimation unit that estimates position information of each mobile object, a route planning unit that creates a route plan of each mobile object based on the estimated position information, an acceleration data acquisition unit that acquires acceleration data acquired from the acceleration sensor, and a mobile object position acquisition unit that acquires an actual position of each mobile object using the position sensor, and learns the deep reinforcement learning model so as to correct a deviation amount between an actual position of the mobile object acquired using the position sensor and an estimated position of the mobile object based on the acquired acceleration data.
1 . A group control system configured to control a plurality of autonomous mobile objects, the system comprising:
an autonomous travel control device comprising a first processor configured to estimate positions of the plurality of autonomous mobile objects at a predetermined timing, and to create route plans based on the estimated positions; and
a group control device comprising a second processor, wherein
the second processor is configured to
receive output from a deep reinforcement learning model and assign route plans to the respective plurality of autonomous mobile objects, the deep reinforcement learning model being trained using (i) acceleration/deceleration data at predetermined time intervals from an acceleration sensor mounted on each of the plurality of autonomous mobile objects while the plurality of autonomous mobile objects are moving, (ii) actual positions of the plurality of autonomous mobile objects from a position sensor mounted on each of the plurality of autonomous mobile objects, and (iii) estimated positions of the plurality of autonomous mobile objects estimated by the first processor, to correct deviation amounts between the actual positions and the estimated positions, and to learn so that the actual positions of the plurality of autonomous mobile objects do not overlap with each other,
manage operation so that the plurality of autonomous mobile objects do not interfere with each other, and
when an environmental change occurs, receive output again from the deep reinforcement learning model and assign the route plans again to the respective plurality of autonomous mobile objects, and
the deep reinforcement learning model is retrained to cause the plurality of autonomous mobile objects not to interfere with each other.
2 . The group control system according to claim 1 , wherein the environmental change is at least one of a layout-change, an increased number of the autonomous mobile objects, and generation of obstacles.