Cooperative V
499
Part B | 22.1
22. Cooperative Vehicle Target Tracking
Frank Ehlers
As autonomous ocean vehicles become more affordable and reliable, applications of multivehicle
teams become more feasible. Cooperative vehicle target tracking is a promising application since
in many sport, military, and biological endeavors cooperative strategies have proven themselves
to be advantageous over noncooperative strategies. The sophistication of winning human team
interactions indicates the challenge inherent in
programming a fleet of autonomous vehicles to
work cooperatively to effectively and efficiently accomplish a goal. For a multivehicle team operating
in the continuous and transient ocean environment, meeting this challenge involves optimizing
a high-dimensional parameter space, even more
so in cases when the target has the ability to make
intelligent choices to avoid being tracked. In this
chapter, recent progress to construct a theoretical
framework and recent applications for maritime
surveillance are presented. A methodology is developed that can help bridge the gap between
22.1 General Theoretical Framework ............ 499
22.1.1 Control and Estimation Approach 501
22.1.2 Multiagent Approach ................. 502
22.1.3 Cellular Automata Approach ....... 502
22.1.4 Probabilistic Graphical Models
Approach .................................. 502
22.2 Distributed Sensing, Control,
and Decisions ...................................... 503
22.3 Multistatic Sonar ................................. 505
22.4 Maritime Surveillance .......................... 510
22.5 Effective Coordination Schemes ............ 511
22.5.1 The Understanding Phase .......... 511
22.5.2 The Trading Methodology .......... 511
22.6 Conclusions and Recommendations ...... 512
References................................................... 513
the top-down view starting with theoretical concepts and the bottom-up view dealing with all
details of real experimentation and execution at
sea.
22.1 General Theoretical Framework
Target tracking in surveillance applications is essentially a problem of connecting signature dots (glimpses
of a target) in a noisy, uncertain environment. Cooperative vehicle target tracking is a method in which
multiple sensing platforms (vehicles) jointly react to
a target’s behavior, which may or may not be evasive, to
maintain contact and extend previously observed tracks.
The maritime environment is noisy, uncertain, and spatially extensive. In maritime surveillance applications
noise and uncertainty can be reduced through persistence and adaptability. For operations over large areas,
the scalability of any target-tracking methodology is an
important practical consideration. Autonomous ocean
vehicles are useful tools for maritime target tracking
since they enable both persistence and adaptive, scalable approaches [22.1, 2].
Teamwork can be efficient and effective in successfully completing difficult tasks. Teamwork is a cooperative effort by a group of individuals or agents to achieve
a common goal, wherein individual team members contribute different skills, knowledge, and interests to the
effectiveness of the group. In a game setting, targets
are modeled to make locally optimal decisions based
on their established tactics (or internal mental state) independently (from their perspective) of the activities of
the surveillance team. Constrained by established tactics, the target’s dynamics can be expressed by a control
algorithm that formulates the objective of the chosen
mental state. For example, if the mental state is to drift
to save energy, the target dynamics can be linked to
the oceans dynamics. The surveillance team’s goal is
to track the target with distributed mobile sensors under
centralized and/or distributed control in a coordination
scheme that may be called synchronized covertness.
In this chapter, the advantages (or even necessity) of
teamwork for target tracking using autonomous ocean
vehicles are investigated. A trading methodology is described which enables the generation of coordination
499
Part B | 22.1
22. Cooperative Vehicle Target Tracking
Frank Ehlers
As autonomous ocean vehicles become more affordable and reliable, applications of multivehicle
teams become more feasible. Cooperative vehicle target tracking is a promising application since
in many sport, military, and biological endeavors cooperative strategies have proven themselves
to be advantageous over noncooperative strategies. The sophistication of winning human team
interactions indicates the challenge inherent in
programming a fleet of autonomous vehicles to
work cooperatively to effectively and efficiently accomplish a goal. For a multivehicle team operating
in the continuous and transient ocean environment, meeting this challenge involves optimizing
a high-dimensional parameter space, even more
so in cases when the target has the ability to make
intelligent choices to avoid being tracked. In this
chapter, recent progress to construct a theoretical
framework and recent applications for maritime
surveillance are presented. A methodology is developed that can help bridge the gap between
22.1 General Theoretical Framework ............ 499
22.1.1 Control and Estimation Approach 501
22.1.2 Multiagent Approach ................. 502
22.1.3 Cellular Automata Approach ....... 502
22.1.4 Probabilistic Graphical Models
Approach .................................. 502
22.2 Distributed Sensing, Control,
and Decisions ...................................... 503
22.3 Multistatic Sonar ................................. 505
22.4 Maritime Surveillance .......................... 510
22.5 Effective Coordination Schemes ............ 511
22.5.1 The Understanding Phase .......... 511
22.5.2 The Trading Methodology .......... 511
22.6 Conclusions and Recommendations ...... 512
References................................................... 513
the top-down view starting with theoretical concepts and the bottom-up view dealing with all
details of real experimentation and execution at
sea.
22.1 General Theoretical Framework
Target tracking in surveillance applications is essentially a problem of connecting signature dots (glimpses
of a target) in a noisy, uncertain environment. Cooperative vehicle target tracking is a method in which
multiple sensing platforms (vehicles) jointly react to
a target’s behavior, which may or may not be evasive, to
maintain contact and extend previously observed tracks.
The maritime environment is noisy, uncertain, and spatially extensive. In maritime surveillance applications
noise and uncertainty can be reduced through persistence and adaptability. For operations over large areas,
the scalability of any target-tracking methodology is an
important practical consideration. Autonomous ocean
vehicles are useful tools for maritime target tracking
since they enable both persistence and adaptive, scalable approaches [22.1, 2].
Teamwork can be efficient and effective in successfully completing difficult tasks. Teamwork is a cooperative effort by a group of individuals or agents to achieve
a common goal, wherein individual team members contribute different skills, knowledge, and interests to the
effectiveness of the group. In a game setting, targets
are modeled to make locally optimal decisions based
on their established tactics (or internal mental state) independently (from their perspective) of the activities of
the surveillance team. Constrained by established tactics, the target’s dynamics can be expressed by a control
algorithm that formulates the objective of the chosen
mental state. For example, if the mental state is to drift
to save energy, the target dynamics can be linked to
the oceans dynamics. The surveillance team’s goal is
to track the target with distributed mobile sensors under
centralized and/or distributed control in a coordination
scheme that may be called synchronized covertness.
In this chapter, the advantages (or even necessity) of
teamwork for target tracking using autonomous ocean
vehicles are investigated. A trading methodology is described which enables the generation of coordination