Odor Localization using Gas Sensor for Mobile Robot Nyayu Latifah Husni Electrical Department State Polytechnic of Sriwijaya Palembang, Indonesia nyayu_latifah@polsri.ac.id Ade Silvia Handayani Electrical Department State Polytechnic of Sriwijaya Palembang, Indonesia ade_silvia@polsri.co.id Siti Nurmaini 2 Robotic and Control Research Lab, Faculty of Computer Science, University of Sriwijaya siti_nurmaini@unsri.ac.id, Irsyadi Yani 3 Mechanical Engineering Department, Faculty of Engineering, University of Sriwijaya yani_irs@yahoo.com Abstract—This paper discusses the odor localization using Fuzzy logic algorithm. The concentrations of the source that is sensed by the gas sensors are used as the inputs of the fuzzy. The output of the Fuzzy logic is used to determine the PWM (Pulse Width Modulation) of driver motors of the robot. The path that the robot should track depends on the PWM of the right and left motors of the robot. When the concentration in the right side of the robot is higher than the middle and the left side, the fuzzy logic will give decision to the robot to move to the right. In that condition, the left motor is in the high speed condition and the right motor is in slow speed condition. Therefore, the robot will move to the right. The experiment was done in a conditioned room using a robot that is equipped with 3 gas sensors. Although the robot is still needed some improvements in accomplishing its task, the result shows that fuzzy algorithms are effective enough in performing odor localization task in mobile robot. Keywords— odor localization; fuzzy logic; TGS. I. INTRODUCTION The development of odor localization research has grown widely and rapidly. Most of the researches used static [1] and dynamic instruments [2], [3], [4]. In static system, the gas sensors were placed in determined spots. Some drawbacks of them occurred, such as ineffectiveness of the sensors due to the working areas of sensors that were limited by the static range between the sensors and the sources. It is contradictive with the dynamic system, i.e. the use of mobile instruments where the gas sensors are integrated to mobile robots or mobile devices. Being integrated to the mobile robots makes the gas sensors be able to reach wider areas. Although the range between the gas sensors and the sources still influence the working areas of mobile sensors, the wide areas can be achieved by mobile characteristics of the robots where gas sensors placed. The localization of odor sources were widely investigated by researchers using simulation [5], [6], [7] or real experiments [8], [9], [10]. The effectiveness of the robots in localizing the sources depends on the methods used. Some of the previous researchers used the algorithms as follows: 1. imitated the behavior of the animals (chemotaxis and anemotaxis) [11], [12], [13]; 2. based on the flow of the fluid (fluxotaxis) [14], [15], [16]; 3. used entropy of the posterior probability field (infotaxis) [17], etc. However, most of the algorithms has drawbacks, such as low search accuracy and efficiency due to their dependence on the wind direction [18]. Jiandong [18] tried to find another way using fuzzy logic to localize the odor. The path where the robot should go was determined by the rate change of the plume sensed by the 3 sensors mounted on it. Fuzzy logic was successful in controlling the trajectory of the robot. Even though, the validation of the fuzzy logic was done only in simulation. It was far from the real one. In this research, a real robot was developed and implemented in a real experiment. The robots were equipped with some gas sensors that have a task to supply the inputs data to the fuzzy logic algorithm. Some other researchers also used fuzzy logic in their experiments of odor localization [19], [20], [21], [22]. Most of them used fuzzy to control the communication network. X. Cui [19] implemented fuzzy logic in swarm robots of mobile sensor networks. It is used to control nodes of the sensor network in determining the next optimal node deployment location. Siti Nurmaini in [20] proposed Fuzzy- Kohonen Networks and Particle Swarm Optimization (FKN- PSO) to localize the odor source. The result was then compared to the Fuzzy-PSO. It showed that FKN-PSO was more efficient than Fuzzy-PSO. Other researches, such as in [21] and [22], show that the fuzzy logic were used as sensor's information processor. P. Jiang [21] used fuzzy logic to process multi inputs of the sensors (olfaction, vision, wind speed/direction, distance and position of robot). More detailed and accurate decisions of these inputs were got easily using Fuzzy logic. The outputs of the fuzzy were set up into six behaviors, including obstacles avoidance, odor source declaration, nearest distance-based visual searching, up-wind searching, path planning, chemotaxis searching, and random searching. The proposed algorithm was successful in increasing the ability of the robots in finding the plume. Siti Nurmaini in [22] was successful using fuzzy logic in finding the best target position of each swarm robots. The fuzzy logic was only activated when the gas sensors were inactive and robots moved in unknown areas. When gas sensors were active, the Fuzzy-PSO was used. Fuzzy-PSO was successful to control the trajectory and movement of the robots. In this research, the information of odor concentration from gas sensors was used to determine the track of the robots. The high concentration indicated that it was the way of the source came from. Therefore, the fuzzy output will be Proc. EECSI 2017, Yogyakarta, Indonesia, 19-21 September 2017 978-1-5386-0549-3/17/$31.00 ©2017 IEEE . 584 brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by Proceeding of the Electrical Engineering Computer Science and Informatics