Issue link: https://resources.mouser.com/i/1442859
Texas Instruments The Future of Robotics | 9 Autonomous Robotic Systems To gain knowledge and respond to their environmental contexts, autonomous robotic systems must do four primary tasks: 1. Sense–Collect and receive data 2. Perceive–Interpret the data, awareness of the autonomous robotic system of itself and its situational context 3. Plan–Plot a course of action 4. Act–Take action on the plan and adjust along the path Artificial intelligence (AI) helps oversee these processes. For AI to fulfill its promise in many industries, the fusion of sensor data happens during inference learning-the part that executes the trained machine learning (ML) algorithm–and must occur in near real- time. This means that ML and deep learning (DL) models must be located on the edge, deploying the inference into an embedded system. Sensor Fusion A signal chain for adaptive, self-learning AI robotic systems requires a fusion of diverse sensor data in real-time. Sensor fusion primarily takes place for an autonomous robotic system at the intersection of sensing and perceiving. Like the human's five senses, sensors in the autonomous robotic system collect data from the outside world. The collected data must be interpreted and turned into knowledge. This means that the various sensors must have their data merged and seamed together to be processed Figure 1: For increased accuracy and robustness under a wide variety of conditions, data from multiple sensor types of complementary modalities need to be fused to view the same scene. Figure 2: A stack of chocolate brownies with mint leaf on top, homemade bakery and dessert excite various human senses. further for the system to obtain a better understanding. This phase is called sensor fusion. Like the human senses, sensor fusion is combining different types of data (Figure 1). The human experience of consuming dessert combines many senses at once, including vision, taste, and smell, to coordinate a pleasant experience. These human senses work together to reduce the uncertainty involved in consumption. It gives us better insights from data (Figure 2). Say you are hungry and pull a dessert out of the refrigerator. It might have a visual look that provides input to the brain that, although this was at one time a potentially yummy dessert, its visual appearance has now shown that it might be suspect. The same might hold for its smell. It might be that the visual and smell senses respond so well to what they are collecting that you know that this dessert is going to be the best you have ever had. Regardless of the context, your various senses work together to help navigate the task at hand and make decisions while simultaneously giving you feedback. Your various senses work together to help navigate the task at hand and make decisions while simultaneously giving you feedback. " "