Avular Autopilot Inference

This page offers an introduction to the Avular’s autonomy stack, called Autopilot Inference. It starts with the motivation to give a robot autonomous capabilities, followed by an overall design of the software functionalities that provide these autonomous capabilities to our robots. The functionalities are then used to explain the principles that are used by the robot when planning and executing its operation. Which will also cover the behavior that can be expected from the robot.

Direct control or autonomous

The robot may be acquired either with or without autonomy. In order to provide our robots with some level of autonomy Avular is developing a software stack called Autopilot Inference. Autonomy of a robot is desired when a user or operator does not want to directly control the robot continuously but instead give high-level commands or tasks, thereby allowing the user to focus his or her attention elsewhere. Depending on the composition of this stack one may expect a higher or lower level of autonomy from the robot. In general, a higher level of autonomy implies that the robot is authorized and able to make proper decisions on how to continue its operation, i.e., the interventions still needed from a user to mitigate any downtime of the robot are limited. A lower level of autonomy implies that the robot has a more strict authorization and thus that the scope in which it is allowed to make its own decisions is limited (either because the robot is not able to make complex decisions, or because a user does not want to give such authority to the robot)..

Info

Please continue to the autonomous capabilities of the Autopilot Inference.