Publikace

This paper describes a novel type of metal markers for eddy current metal detection systems using precise positioning information. The new design for metal markers which are compounded by an ordered array of metal plates with different magnetic parameters has three fundamental advantages: 1. the markers are sensed directly by the metal detector, and no additional hardware is required; 2. the signatures are sharp; 3. not only the position but also the speed of the detector head can be evaluated. In the case of a stand-alone inertial navigation system for evaluating the position of the detector, the precision would be low due to the sensor noise and its integration. However, knowledge of the relative positions and potentially of the detector head speed over the markers can be used as auxiliary information, and so the eddy-current mapping system will be capable of determining the size of the magnetic imprint of a detected metal object and its position relative to the markers. This increases the reliability of the object discrimination during humanitarian demining, and decreases the numbers of false alarms, which nowadays constitute 99.9% of all alarms emitted by an ordinary mine detector.

This paper deals with a cost effective inertial reference unit design providing both MEMS based navigation unit calibration means and an attitude and heading measurement system in a directional gyro mode of operation. A main contribution of this paper is a novel design of such a universal system not primary relying on a magnetometer (MAG) or GPS aiding. Generally, without this aiding Attitude and Heading Reference Systems (AHRSs) are not directionally stable. Also, having precise reference in a calibration process is crucial and in most cases the solution is expensive. We thus replaced an expensive solution of both applications with the cost effective one suiting mentioned purposes using only one single axis fiber optic gyro supported by an inertial MEMS based aiding system. We also proposed a calibration procedure and blended solution to provide both stable and reliable navigation data with accuracy better than 5 deg/h specified by the TSO-C5e.

Data processing of low-cost inertial sensors has been and still is a big issue for many engineers. Inertial sensors compose a measurement unit used as the basis of navigation systems. These systems can be found almost everywhere; it does not matter if it is in pharmaceutics, robotics, automotive industry, aerospace etc. However, there is always a need to specify under which conditions the systems are supposed to operate. This paper deals with the processing of inertial sensor data measured in an environment with strong vibrations on the ultra-light aircraft ATEC 321. It was a great challenge to process the data in such a way as to ensure the acceptable behavior of the artificial horizon system. Because the system should rely only on acceleration and angular rate sensing, the correctness and efficiency of data filtering as well as consecutive data processing play a key role. Results regarding data analyses and comparisons of different approaches are presented.

This paper describes the results of comparison analysis regarding three algorithms for heading determination using low-cost inertial sensors: the stand-alone strapdown approach, magnetometer aiding approach, and strapdown approach with errors modelled, estimated and compensated using Kalman filtering. Experimental testing with precise rotational tilt platform was performed. Error analysis was carried out to qualitatively evaluate the best approach intended for implementation into unmanned aerial systems for navigation purposes.

Distinguishing of mines from scrap metal is a difficult task that provides challenge when seeking an ultimate solution. An eddy-current imaging feature added to standard handheld metal detector is an option for such discrimination. Our metal detector was enhanced by an inertial measurement unit aided by velocity information, measured by optic flow sensor. It lead to further improvement of the proposed procedure by means of position and orientation tracking. The achieved accuracy of mine signal localization is 1 cm in 1x1m scanning area.