<title>Distributed tactical reasoning framework for intelligent vehicles</title>

Abstract

In independent vehicle concepts for the Automated Highway System (AHS), the ability to make competent tacticallevel decisions in real-time is crucial. Traditional approaches to tactical reasoning typically involve the implementation of large monolithic systems, such as decision trees or finite state machines. However, as the complexity of the environment grows, the unforeseen interactions between components can make modifications to such systems very challenging. For example, changing an overtaking behavior may require several, non-local changes to car-following, lane changing and gap acceptance rules. This paper presents a distributed solution to the problem. PolySAPIENT consists of a collection of autonomous modules (termed "reasoning objects"), each specializing in a particular aspect of the driving task -classified by traffic entities rather than tactical behavior. Thus, the influence of the vehicle ahead on the available actions is managed by one reasoning object, while the implications of an approaching exit are managed by another. The independent recommendations from these reasoning objects are expressed in the form of votes and vetos over a "tactical action space", and are resolved by a voting arbiter. This local independence enables PolySAPIENT reasoning objects to be developed independently, using a heterogeneous implementation (e.g., obstacle avoidance modules can use potential fields while exit reasoning is performed using fuzzy rules). PolySAPIENT vehicles are implemented in the SHIVA tactical highway simulator, whose vehicles are based on the Carnegie Mellon Navlab robots.

Control-driven mapping and planning

“…If controllers' voting weights are not properly balanced, one controller may dominate the arbitration, either preventing the robot from making progress to higher-level goals or allowing the robot into undesirable states. Because this weighting is typically an empirical process and dependent on implementation and environment, we have added robustness in a manner similar to [11] by supplementing the voting scheme with "vetoes". Each controller, in addition to its allotment of votes is given the option to veto each of the available actions.…”

Learning from examples in unstructured, outdoor environments

“…If the behavior's voting weights are not properly balanced, one behavior's input may dominate the tally, either preventing the robot from achieving higher-level goals or allowing the robot to enter an undesirable state. Noting that this weighting is typically an empirical process and dependent on both implementation and the robot's environment, an additional layer of robustness has been added by supplementing the voting scheme with "vetoes" (R. Sukthankar, 1997). Each behavior, in addition to getting an allotment of votes to apply to the action set, is given the opportunity to veto each action in the set.…”

Expert system design for an Autonomous Driver Evaluation System

“…Sukthankar et al [2] presented a distributed solution, which consists of a collection of reasoning objects that vote upon a set of possible actions for dealing with the complexity of tactical reasoning. Rosa et al [3] implemented a fuzzy expert system to organize traffic regulations in a town area.…”