Landmark-based Navigation of an Unmanned Ground Vehicle (UGV).
Scientific Publication
- Report Number:
- DSTO-TR-2260
- Authors:
- Kim, J.; Hmam, H.
- Issue Date:
- 2009-03
- AR Number:
- AR-014-415
- Classification:
- Unclassified
- Report Type:
- Technical Report
- Division:
- Weapon Systems Division (WSD)
- Release Authority:
- Chief, Weapons Systems Division
- Task Sponsor:
- CWSD
- Task Number:
- 07/249
- File Number:
- 2008/1126788
- Pages:
- 37
- References:
- 8
- Terms:
- Navigation; Optimisation
- URI:
- http://hdl.handle.net/1947/9896
Abstract
We present a method for estimating the position and orientation of a ground vehicle in an environment with landmarks. From the geometric relationships, we derive a set of linear equations with a quadratic constraint, which forms the basis for our optimisation problem. We also extend the problem to associating two sets of measurements taken at two successive locations to improve the navigation accuracy. This method is efficient and the performance is robust against large measurement errors.
Executive Summary
We present an efficient method for localising an unmanned ground vehicle (UGV) in a two dimensional environment with landmarks. We assume that the vehicle can identify these landmarks and measure their bearings. We derive a set of linear equations that relate the UGV’s position and orientation to the measured bearings, assuming perfect measurements. Having noise in the measurements, we establish an estimator, based on minimising a constrained quadratic cost function. Landmark-based navigation is useful where GPS is not adequate because of signal blockage and interference. Possible defence applications include robots performing the following tasks: indoor bomb search, landmine location, autonomous surveillance of military bases or nuclear sites, and logistics in urban battle zones. The proposed algorithm is efficient and the performance is consistent given that the UGV is inside the boundary formed by the landmarks. The estimation accuracy is shown to improve as the number of landmarks increases. Given that the number of landmarks is limited, we extend the initial formulation and propose another method that utilises two sets of measurements from two locations and estimates the current position and azimuth with significantly improved accuracy. The performance is robust against large measurement errors, and can be coupled with an inertial navigation system to enable higher rate navigation with greater accuracy, especially in outdoor settings.