INTRODUCTION

Guidance of agricultural vehicles has received the attention of researchers for nearly 50 years. From a review of the literature, several motivations are apparent from the effort that has been expended towards the development of a guidance system, be it autonomous or partially autonomous. The need to relieve the operator of continuously making steering adjustments while maintaining the attached or towed implement at some level of acceptable performance is perhaps the most frequently cited reason for the need for a guidance system (Wilson, 2000). Research activities concerning automatic guidance of agricultural vehicles - or implements - have led to various solutions. Sensors, including mechanical ones, Global Navigation Satellite Systems (GNSS), machine vision, laser triangulation, ultrasonic and geomagnetic, generate position, attitude and direction-of-movement information to supply control algorithms. Actuators, like hydraulic valves, are used to transform guidance information into changes of position and direction. In the European Union, depending on who is funding the project, the systems range from a PC, with a frame grabber or a GNSS receiver used to guide an implement along a predefined path with speeds up to 3 m sec, to a multiprocessor bifocal road recognition system for autonomous cars driving on motorways with a speed of 130 km h^-1 (Keicher and Seufert, 2000).

In mechanized agriculture applying most of technologies such as new varieties of seeds, new pesticides, modern equipment, etc. needs to use tractors in farm as power resource. Precision guidance of the tractor in the field is critical factor because it affects fuel consumption, time used for completing a specific operation and amount of chemical which are used. Tractors in each path cover a limited area bigger than width of tractor so doing subsequent passes accurately and maintaining in a straight line (which is necessary for using successive agricultural machinery) is a big challenge for tractor driver. Because he should control work of implement and concentrate on guidance of tractor in a straight line simultaneously (Tillett, 1990).

Due to long hours of working in agricultural operations and high level of stresses, tractor driver becomes tired soon and makes a little overlapping about 5% in each path. Eventually fuel and lubricants consumption would increase that cause more air pollution, likewise more applying of chemicals and more pollution of soil and underground waters and, of course, increase of time needed to complete an specific operation which causes not completing it in optimum time period and decreases yield. Beside these by using TDAS tractor's driver fatigue would be decreased and he can do his duties during the day more efficiently. Some common technology used for navigation such as GPS. DGPS, RTKDGPS are not applicable for Iranian farmers. High investment requirement and complexity of technology and small size of farms eliminate use of them. In past years many effort has been expended towards the development of a guidance system in two categories, full autonomous or partially autonomous. Studies by Kaminaka et al. (1981) have shown that steering accuracy decreases rather dramatically as extra demands are placed on the operator. Furthermore, such studies have tended to confirm the observed fact that increased field speeds require the operator to be more attentive to the driving function, often at the expense of reduced equipment performance.

In Hokkaido University, a crop row detector equipped with a one-dimensional image sensor was developed for use in crop husbandry machinery. The principle was that the crop row image was converted, in the hardware, to a one-dimensional gray scale level signal and the software estimated the offset and heading errors. The execution time was about 40 msec and the accuracy was sufficient for practical use when the camera offset was up to 0.10 m and the attitude angle not more than 6°. Finally, sensor performance tests were conducted on the following of the crop row (Torii, 2000).

The purpose of this research was to develop a Tractor's Driver Assistant System (TDAS) as a modern technology for reducing his stresses and fatigue, save time and money and help to achieve sustainable agriculture with respect to social and economical situations of Iran as a developing country in which farmers do not have enough money to buy modern technologies from developed countries and, of course, sometimes buying them due to political conditions is impossible.

In this study first, the methods of aiding system of agricultural machines are explained and then the automatic system with its capabilities in MF285 tractors is proposed.

Debain et al. (2000) developed a guidance-assistance system for agricultural machines. It is based on analysis of the vehicle’s environment by image processing to deduce a control law in the image space. The results of this study the feasibility of such a project and the necessity to know the reliability of the image processing results to secure the viability of the complete system.

A guidance-assistance system for agricultural vehicles.

However, it is interesting to note that all approaches revert to basically two distinct modes:

Typical systems that have been developed in the first mode include:

All of the mentioned systems have disadvantages that restricted their application, such as high capital investment, not mobility or complexity. For instance, the cost of buried cables as used in the second mode is difficult to justify on large fields that are not intensively farmed. Radio beacons can be portable and allow for changes in the direction of field operations. However, the cost of signal processing plus line of sight limitations appears to have limited the application of this approach to special situations where some of the characteristics, e.g., ability to operate at night or under fog conditions, can be exploited (Reid et al., 2000). In beside of these systems, there are modern technologies such as GPS, DGPS, RTK GPS, GDS machine vision and fusion sensors, which have some basic requirements for technology, investment and skill.

Typical studies reported in the second mode in some researches two wheels attached to mechanical linkages pin-connected to the tractor straddled the marker furrow, Harries and Ambler (1981) mounted optical sensors for non contact sensing of the furrow wall. Connor et al. (1996) developed a steering controller based on a set of linear motion equations. Automation requires a steering controller with stable and fast response. The design of steering controllers should compensate for non-linearity and many unknown factors involved in steering agricultural equipment.

The common feature in these approaches is that the measurement (or on-off indication) is the lateral distance (or offset) from a point on the tractor to the furrow being followed. In general, the various sensors used provided an indication of current offset.

But the most important advantages of these systems are simple construction, low cost and mobility. In this research we focused on developing late systems for most popular agricultural tractor in Iran, MF285 as appropriate technology because it has characteristics which are matched with definitions of appropriate technology such as: