The Robot Intelligence Kernel

The Robot Intelligence Kernel (RIK) is a portable, reconfigurable suite of perceptual, behavioral, and cognitive capabilities that can be used across many different platforms, environments, and tasks. The RIK coupled with a virtual 3D interface have been shown to dramatically improve human-robot interactions across a variety of navigation and exploration tasks. The Robot Intelligence Kernel is a portable, reconfigurable suite of perceptual, behavioral, and cognitive capabilities that can be used across many different platforms, environments, and tasks. The RIK integrates algorithms and hardware for perception, world-modeling, adaptive communication, dynamic tasking, and autonomous behaviors in navigation, search, and detection tasks. The integration of software algorithms and hardware takes place over four levels of the RIK. The foundation layer is the Generic Robot Architecture that provides an object-oriented framework and an application programming interface (API) that allows different platforms, sensors, and actuators to interface with the RIK. The second layer is the Generic Robot Abstractions layer which takes data from the first layer and abstracts the data so that it can be used in algorithms that are designed for generic robot systems. The third layer is comprised of many reactive and deliberative robot behaviors that take, as input, the abstractions from the second layer and output commands for the robot to follow. The fourth and final layer provides the “Cognitive Glue” that orchestrates the asynchronous firings of multiple behaviors towards specific applicationbased tasks. The goal of the RIK has been to create an integrated suite of primitive behaviors that will degrade gracefully in the face of uncertainty. Instead of crafting algorithms to be optimal, these behaviors are crafted to be robust, responsive, and adjustable across different search, detection, and exploration tasks. These behaviors are intrinsic, meaning that they can provide a basic level of competency to the robot even when all external input (such as global positioning (GPS) or communication with a human operator) is unavailable. The RIK is not just a collection of independent, unrelated behaviors, but rather an integrated suite of capabilities orCopyright c © 2006, American Association for Artificial Intelligence (www.aaai.org). All rights reserved. Figure 1: The Virtual 3D Display shows semantic entities added by the robot or the human. chestrated by the cognitive glue. Hence, the RIK is a collection of interdependent meta-level capabilities providing a) real-time map-building and positioning, b) reactive obstacle avoidance and path planning, c) high-speed waypoint navigation with or without GPS, d) self-status awareness and health monitoring, e) online adaptation to sensor failure, f) real-time visualand laser-based change detection, g) human presence detection and tracking, and h) adaptive, multimodal communication, and i) problem solving that can be used to accomplish high-level tasking, dynamic autonomy, and dialoging for flexible human-robot teaming. True teamwork requires a shared understanding of the environment and task. To support a dynamic sharing of roles and responsibilities between a human and a robot, the RIK employs a representation that allows both the human and robot to reason spatially about the world and to understand the spatial perspective of the other. Understanding the robot’s perspective of the environment allows the human to predict robot behavior. Understanding the human’s perspective enables the robot to interpret and infer intentions from tasks given by the human. Through sensor fusion and probabilistic reasoning (Konolige 2004), the robot creates a 3D, virtual representation of