Car Navigation and Collision Avoidance System with Fuzzy Logic Andri Riid, Dmitri Pahhomov and Ennu Rüstern Department of Computer Control Tallinn University of Technology Ehitajate tee 5, 19086 Tallinn, Estonia E-mail: andri@dcc.ttu.ee Abstract—A navigation control and collision avoidance system for delivering a car to the arbitrarily positioned loading dock is designed, based on the fuzzy trajectory mapping unit (TMU). Simulated driving experiments in different environmental conditions demonstrate that the designed system shows good performance. Modular structure of the control system facilitates both efficient control knowledge acquisition (which is encapsulated in TMU) as well as further development of the control system to accomplish more demanding tasks. I. INTRODUCTION In fuzzy control research, the cab part of Nguyen and Widrow’s truck backer-upper [1] is often picked as a test object [2-8] for its nonlinearity on one hand and lack of traditional control system design methods that could be applied for this problem, on the other. At first glance, the control problem also seems one of these cases where the historical application area of fuzzy controllers - knowledge- based control - would be appropriate solution. This seems so because ability to drive a car is a very common skill among people thus it should not be too difficult to find an expert whose verbal instructions would then constitute the core of the control system. Car driving skill, however, is usually learned to a degree where it rarely intrudes on consciousness (the occasions when it does are unusual circumstances like a potential accident or a situation the driver is not used to (i.e. he/she has not yet learnt it). Consequently it is difficult to extract appropriate rules from an expert because of one’s inability to explain how one accomplishes the task, and consequent difficulties in putting it down in terms of fuzzy logic. The design of knowledge-based controller therefore becomes much more difficult than was assumed in the first place. In our previous study [9] we have shown how the decomposition of the control problem can substantially simplify and make very natural the most critical part of controller design - expert knowledge acquisition. This is because we become are focused on information concerning car optimal orientation in two-dimensional space (that is much easier to explain than the very actions on the steering wheel), which ultimately results in an efficient solution of the problem. Moreover, the decomposition reduces dimensionality of the problem (which allows us to involve ourselves with more complex control goals such as knowledge-based control solution of the original truck backer-upper problem, as shown in [10]). The present paper further develops the approach taken in [9] and [10], by first generalizing functionality of the central unit of the control system, which allows us to specify an arbitrarily chosen loading dock position and then utilizing it for trajectory management in subsequently developed collision avoidance system. II. CONTROL OBJECT The car position is determined by three state variables x, y and Φ = [-90°, 270°], where the latter is the angle between truck's onward direction and the x-axis (Fig. 1). The width and length of the car are denoted by w and l, respectively (w = 2, l = 4). Φ f 180° 90° 270° 0° (x f , y f ) (x, y) θ Φ y x Fig. 1. Car and main variables The problem is formulated as follows: the car must arrive from the arbitrary initial position (x i , y i , Φ i ) to the predefined loading dock (x f , y f , Φ f ). Car moves forward or backward with the fixed speed (i.e. speed control is not our concern). To control the car, appropriate steering angle θ = [-45°, 45°] must be provided. Moreover, the car must be able to avoid contacts with the bounding walls (ranges of x and y can be freely specified) and with any other obstacles on its way.