Fire Fighting Mobile Robot
KePIX is a fire fighting mobile robot. Based on previous research, KePIX used LDR and phototransistor sensor. Robot controlled by 2 mode, on-off and fuzzy control. On-off method used for detecting existence of fire and it's distance. Fuzzy control used for approaching point of fire. Robot controller used microcontroller 89C52 (Hindriyanto, 2005).
On the next research,there was additional system for mobile robot, that are serial data communication system between mobile robot model and personal computer; and distance measurement system using ultrasonic sensor (Pradhono, 2005).
Generally, mobile robot has hardware and software. The hardware section formed by distance measurement circuits, infrared sensor circuits, control motor for velocity circuits, steering wheel angle control circuits, fire extinguishes control circuits, and interface circuits with microcontroller. And the software section compiled by data acquisition function and navigation control function.
Fuzzy Logic Control
There is two fundamental reason why I use fuzzy logic as a control system. First, it is very difficult to explain the complexity of the real world accurately, therefor we need approximation. And the other, there is significant knowledge that can not be substitute by machine in the engineering world. Obviously it will great if we can combined human knowledge ability and machine ability backup with mathematics model, control, and sensor measurement. Therefor needed a systematic design concept that can formulated human knowledge and embedded into machine system.
There is four main section on fuzzy-based control system, that are fuzzyfication, fuzzy rule base, inference engine, and defuzzyfication as shown below.
Fuzzy Control System based Mobile Robot Navigation
To build navigation control for mobot (mobile robot), it may accomplish by developing fuzzy control system based motion behavior. In this experiment I developed 2 fuzzy control system based behavior, that are obstacle avoidance behavior and goal seeking behaviour. Every each behavior activate with full acces for all sensor reading and control calculation process to result action control. Final command execute based on behavior priority. Every behavior represent the task that mobot must do, informing data sensor and target point. Obstacle avoidance behavior have higher priority than goal seeking behavior.
OBSFIS : Obstacle Avoidance Behavior-based Fuzzy Inference System
OBSFIS activated if object distance detected under 15 cm by ultrasonic sensors. OBSFIS used data from ultrasonic sensors to develop fuzzy membership function that represent distance between mobile robot and surrounding objects. Fuzzy inputs comes from left, front, and right ultrasonic sensor membership function. Membership function for each ultrasonic sensor divided into 3 membership function represent near distance (Zero=ZE), medium distance (Positive Medium=PM), and long distance (Positive Big=PB). Fuzzy outputs comes to steering wheel angle value and velocity. Membership function for steering wheel angle divided into 5 membership function represent positive big (PB), positive medium (PM), zero (ZE), negative medium (NM), and negative big (NB) direction angle. Meanwhile membership function for velocity divided into 3 membership function represent low speed (ZE), medium speed (PM), and high speed (PB).
GOALFIS : Goal Seeking Behavior-based Fuzzy Inference System
GOALFIS activated if object distance detected more than 15 cm by ultrasonic sensors. Fuzzy inputs comes from target distance and target angle membership function. For target distance and angle distance measurement calculated by mapping mobile robot environment with cartesian coordinate. The process, saving actual position and orientation mobile robot value into memory. Initialization to determine coordinate of target position, determine coordinate of initial position of the mobile robot, and determine initial orientation between mobile robot and coordinate frame (β
and angle target toward x axis. The value of angle target (β) calculated as shown below.
The value of target orientation relative toward mobile robot calculated the difference between target angle (
Fuzzy output comes to steering wheel angle. Membership function for steering wheel angle divided into 5 membership function r