سامانه‌ي خودكار توليد نقشه‌ي معناييِ محيط براساس مدل سلسله‌مراتبي در ربات خودگردان با استفاده از اكتشاف مبتني بر هيستوگرام

MOBILE ROBOT SEMANTIC MAPPING USING HISTOGRAM BASED FRONTIER EXPLORATION

«نقشه‌ي معنايي » ابزاري مناسب براي ايجادِ تواناييِ استنتاج در ربات هاي خودگردان است. در اين نوشتار سامانه‌يي يك‌پارچه و هوشمند براي توليد خودكار نقشه‌ي معناييِ سلسله‌مراتبي از محيط واقعي پيشنهاد شده است. در اين سامانه، روشي نوين براي دسته بندي ويژگي هاي ديداري عمومي و محليِ مناطقِ محيط معرفي شده است كه مبتني بر حافظه‌ي كوتاه‌مدت بوده و از آن به‌منظور مكان يابي معنايي ربات ــ در زمان واقعي ــ در فرايند توليد نقشه‌ي معنايي استفاده شده است. به‌منظور توليد نقشه‌ي معنايي از تمام بخش هاي محيط، واحد كنترلي در اين سامانه از روشِ پيشنهاديِ كم‌پردازشي براي اكتشاف خودكارِ محيط بهره مي گيرد. اين واحد براساس هيستوگرام مناطق ناشناخته، مكان هدف آتي را تعيين و با توليد كوتاه ترين مسير ايمن، ربات را به آن محل هدايت مي كند. نتايج پياده سازي سامانه ‌ي پيشنهادي در يك محيط واقعي، بيان‌گر توليد نقشه‌ي معنايي از تمام بخش هاي محيط با صحت 90درصد است.

Home and service robotics is an area of intelligent robot applications. In this regard, robots need to collect sufficient perceptional information from their environment in order to decide how to perform their tasks. The environment semantic map is a helpful base for robot decisions. In this paper, a system is proposed to generate a semantic map of the environment. In this system, a place classification method is used, in order to classify the images into a set of predefined classes. When the region of the current place of the robot is recognized, the new information of the region is correlated to other components in the semantic map. The generated semantic map consists of the room shape model, the metric and topological map and the appearance model of each region. Also, the proposed system benefits from a low computational histogram based exploration method, which suggests navigating the robot towards the boundaries between free and unknown areas in the map in order to facilitate autonomous semantic map generation. A global occupancy grid map of the environment is constantly updated, based on which, a global frontier map is calculated. Then, a histogram based approach is adopted to cluster frontier cells and score these clusters, based on their distance from the robot, as well as the number of frontier cells they contain. In each stage of the algorithm, a sub–goal is set for the robot to navigate. A combination of distance transform and A* search algorithms is utilized to generate a plausible path towards the sub-goal through the free space. In this way, a reliable distance from the obstacles is guaranteed, while searching for the shortest path towards the sub-goal. The whole process is iterated until no unexplored area remains in the map and, consequently, a semantic map of all parts of the environment is generated. Experimental results of the proposed method in the simulated and real environments show that the semantic map of the environment is generated with high accuracy in a short time.