Adaptive collaborative control

Abstract: : This report covers the first year of research and development for the Intelligent Multi-UxV Planner with Adaptive Collaborative/Control Technologies (IMPACT) program. The goal of IMPACT is to integrate autonomous technologies like cooperative control algorithms, intelligent agents, and autonomics frameworks with human-autonomy interfaces to support effective supervisory control of multiple heterogeneous unmanned vehicles (UxVs) in dynamic operating environments. With IMPACT, the human-autonomy team controls multiple UxVs by utilizing high level commands called plays. To evaluate the year one IMPACT system, seven participants were asked to use IMPACT to manage six UxVs in the context of a base defense mission. Participants rated both the overall IMPACT system and system subcomponents (play calling, autonomy, feedback, and test bed) positively. Objective data was also collected on the modality (mouse/keyboard, touchscreen, speech recognition) participants used to call plays. Participants not only used the mouse/keyboard more frequently, but also were faster and more accurate with the mouse/keyboard as compared to the touchscreen and speech recognition. Participant feedback has directly led to planned enhancements for the Spiral 2 IMPACT system.

Abstract: A Command and Control (C2) display system using the Microsoft HoloLens and the Intelligent Multi-UxV Planner with Adaptive Collaborative Control Technologies (IMPACT) has been developed as a demonstration of a new advanced user interface. This allows for human-to-human-to-machine collaboration for situational awareness, decision making, and C2 planning and execution of simulated multi-unmanned heterogeneous autonomous vehicles. The advanced user interface allows multiple operators to collaborate across a shared holographic sand table and control multiple vehicles. Multiple networking frameworks were used to offload the computation of vehicle autonomy and planning algorithms to allow the HoloLens to run efficiently for an improved user experience. Additionally, the concept of pseudo-classified information filtering allows for tiers of classification levels for each HoloLens user derived from a ‘need-to-know’ classification basis.

Abstract: The invention provides a traffic signal time assignment method based on collaborative optimization. The relevance between intersections is determined according to actual distribution of comprehensivesignal lamps and traffic flow, a signal collaborative control area based on SCAN cluster partition is determined, connected intersections with stronger relevance are gathered in the same cluster, andby means of a Boltzmann selection strategy, adaptive collaborative control is performed after regional learning intellectual bodies are subjected to sufficient experience accumulation until the signalcontrol ends, and passing rate of vehicles in a small-area range is increased, so that the traffic efficiency of the overall road network is improved.