Engineering Services
-
Subject Matter Experts for Consultation/Collaboration
-
Field Testing Unmanned Vehicles
-
Sensor Processing/Sensor trade Studies
-
Route Planning/Modeling
-
Obstacle Avoidance
-
Dynamic Route Modeling
-
Leader/Follower Applications
-
Watercraft Control
-
Command and Control Systems Development
-
Cooperative Control for Manned and Unmanned Systems
-
Mapping & Localization (with and without GPS)
-
Custom Applications for Computer Programming and Software Development
Products
Terrain Aware Coordination Toolbox for Intelligent Control (TACTIC)
To be effective, robots need adaptive and flexible behaviors. They must recognize dynamic changes in the environment and modify or regenerate their plans in real-time in order to achieve their overall goals. For use in cluttered environments, planners generate a large number of alternative actions that achieve the goal, evaluate the cost of each alternative and select the one with the lowest cost. This process requires the following steps:
-
Make a graph connecting alternative states
-
Calculate the cost of traverse from one state to another
-
Search the graph to find the least costly concatenation of states that connects the starting state and the ending state
TACTIC Toolbox
Robotic Research provided Demo III and other subjugate programs with obstacle avoidance and autonomous mobility algorithms. We developed Terrain Aware Coordination Tools for Intelligent Control (TACTIC) as a toolbox of algorithms designed to address the three components of planning. This toolbox provides the building blocks necessary to construct and implement tactically aware plans for unmanned vehicles.
-
Search Engines
TACTIC has a multi-heap dynamic A* search engine that, in our experience for autonomous applications, out-performs any other search engine. This search engine is currently being used for powering the latest generation of autonomous mobility planners.
-
Graph generation algorithms
TACTIC has a variety of algorithms that can be used to generate graphs for different uses:
-
Grids and digraphs
-
Just-in-time generated/pruned graphs for large search spaces
-
Kinematically and dynamically relevant graphs (ego-graphs) for autonomous mobility
-
High dimensionality graphs for multiple vehicle coordination
-
Graph fields for multi-level graph generation and execution
-
Cost evaluation
TACTIC includes a set of cost evaluation methodologies that will address tactical behaviors. One important feature for the cost evaluation for tactical behaviors is the creation of probabilistic point clouds that represent and predict the location of an “enemy” as well as the location of other team members. We have developed an innovative approach for creating these point clouds that we believe will revolutionize the cost evaluation for tactical purposes. These tools, in combination with our line of sight tools, can be used to perform probabilistic cost evaluations and consequently populate the world model with new tactical layers. Some of the tools in TACTIC for cost evaluations include:
-
Probabilistic location point cloud algorithms
-
Fast probabilistic line of sight computations
-
Tools for line of sight storage
-
Dynamic interpretation of enemy locations into probabilistic point clouds
-
Proven terrain traversability cost evaluation
-
Tools for registration of terrain information
-
Apriori terrain analysis
-
Sensed data
Probabilistic Red Team Location
Probabilistic Red Team Line of Sight and Cresting
Blue and Red team simulation and planning can be performed using TACTIC
TRAFFIC
Traffic is a software library designed from the ground up to provide three common components for usage in ground robotic systems:
A multidimensional search engine (x,y,time)
Cost computation that adheres to U.S. road rules
On the fly graph generation to create kino-dynamic complainant search space.
Traffic provides libraries for human avoidance, and vehicle avoidance and is utilized by a variety of ground robotic platforms to navigate on road scenarios.
-
Traffic running on a Robotic Ford Escape
Contact us for pricing information
|
