Hardware and software development of the Anglescan tracking system

In this thesis, the author describes all the developmental work of an automated surveying tracking theodolite referred to as The Anglescan Tracking System. This system consists of an optical system, an electronic system (with detailed schematics and printed circuit artwork, included) the software system which is approximately 100 pages of C source code, system identification and controller simulation, modelling and design, and finally instrument calibration and testing effort over a 2 year period. The project is a collaboration between the Department of Electrical and Electronic Engineering and the School of Surveying at the University of Tasmania. It is supported by an ARC university research fund granted to the School of Surveying over a two year period. The initial investigation into the feasibility of the project was carried out by Dr. A. Sprent, then at the North East London Polytechnic (Sprent, 1.12). During this time the optical system of the instrument was developed. Following this investigation, it was decided to further undertake the continuation of the project as a Master's Degree in order to develop the electronics, measurement, tracking and control software. The Anglescan system is similar to a standard surveying theodolite mounted on a tripod and capable of determining the position of a retro-reflective mirror based within the instrument's field of view with respect with its optical centre of field (in both azimuth and elevation planes). This involves a new and different technique for target location than the currently existing electronic theodolites (see sections 1.5, 1.6). One of the main advantages of this method is the improved target locating accuracy, it has been field tested with an accuracy of +1-1 second of arc, and simplicity in optical design. The disadvantages however include increased complexity of both the hardware (electronics), software, and calibration involved. The instrument deviates from a standard theodolite in that it has been fitted with precision DC servomotor drives on the azimuth and elevation axes for target tracking. The servomotors also include optical rotary encoders for measurement.. While tracking, the system operates by continuously locating the target mirror with respect to its centre of field of view, then applying a feedback control signal to the azimuth and elevation motors in order to re-centre the instrument directly onto the target, this is updated at a rate of 30 samples/second. Part of the objectives of this work is to design the necessary hardware and software to achieve this task, this includes the design of a return (laser) analog signal processor card, the design of a target position measurement card, and azimuth/elevation servo motors interface cards (azimuth card and elevation card). Furthermore, the project also requires the design of an appropriate target acquisition and control loop strategy to allow smooth and accurate tracking of a moving target. A number of control system designs have been investigated, including proportional control, pole placement design and linear quadratic control. We have also used system identification techniques to obtain an accurate model of the plant which is necessary for controller design and simulation. To achieve the high precision required in measurement, it is necessary to calibrate the instrument. The calibration method is a complex procedure which uses least squares techniques to produce the required calibration accuracy (this is discussed in detail in section 5). The instrument finds applications in areas such as monitoring earth movements, motion of large structures such as building and bridges, locating earth working machinery such as tractors, bulldozers, graders, concrete and asphalt paving machines where an accurate level surface is required. Another application of particular interest is in surface contour mapping where we require to map a 3 dimensional surface (obtain the surface contour or topography), this application is a slightly more complex problem requiring the use of two instruments, this then becomes a multivariable control problem. As a result of this research work, one paper was presented at the IREECON -91 conference in Sydney and is published in the proceedings [Dirita 1.13]. Another paper was presented at the CONTROL-92 conference held in Perth in November 1992 and is published in the proceedings [Dirita 4.4]