One of the often asked topics in military and industrial transport is the autonomous convoy. Every time soldiers transporting supplies are killed by land mines or ambushes in Afghanistan, the topic of autonomous convoys looms larger, since it could potentially prevent this loss of life.
Because of our expertise in navigation, communication, localization, computer vision and mechatronic vehicle enhancement, we are in a strong position to develop autonomous convoy solutions for your application. We have been working for over two years to integrate our know-how towards an autonomous convoy solution. We have successfully implemented a GPS-based autonomous convoy method using our CHRYSOR platform.
The leader vehicle uses our localization module to create virtual breadcrumbs, which are communicated to the follower vehicles. The unmanned followers then use the localization module as well to reorient themselves relative to the virtual breadcrumbs (see figure 1).
Fig. 1: The leader vehicle communicates virtual breadcrumbs to the follower vehicle.
Afterwards, our path planning method finds an obstacle-free path as consistent with the breadcrumbs as possible. The control module then uses the Device Abstraction Layer to interface to the robot and direct it along the path. That, way the follower vehicles can follow the leader vehicle even when they have to deviate from the breadcrumbs-route to avoid an obstacle (see figure 2). The localization module ensures that the effects of GPS outages and discrepancies between the two vehicles are minimized.
Fig. 2: Follower vehicle follows the path of the leader vehicle, while avoiding obstacles.
We are currently working on a visual system using modern image analysis algorithms to recognize physical color-marked breadcrumbs left by the leader (see figure 3).
Fig. 3: Field test in which the follower vehicle reorients itself through physical color-marked breadcrumbs by using a localization method based on statistical sensor fusion.