Can Humans Teach Robots To Think Like Us?
Although robots are more than capable today of carrying out all kinds of business tasks efficiently and accurately, the concept of building machines that can think like humans has always been a dream for tech companies and smart city developers. However, the actual way in which the human mind works and processes information is up for debate, with several parties having conflicting opinions regarding the same. Once enough data is generated, simulation models can be created to build software that can think along the same rational or emotional lines as humans. Human thinking is generally influenced by a variety of factors—cognitive, behavioral, geometric, kinematic and physical. Using cognitive modeling, such factors can be considered while attempting to create robots that think and behave like humans.
The concept of human thinking is still too vague to be accurately replicated in robots. Even then, multiple types of approaches could be taken to reach the ideal end result—enabling AI and robotic tools to think like humans.
Thinking in humans is deeply influenced by the sensory perception of their immediate environment. Therefore, improving robotic perception can be a significant part of enabling machines to think like humans.
Most modern robots already contain a variety of sensors to improve their data collection and autonomy. For instance, service robots in the fields such as healthcare, agriculture, and entertainment are capable of performing a series of tasks. For mobile robots in such fields, in-built visual sensors and actuators improve their mapping and exploration of different areas. While visual sensors are used for mobility, navigation and visual data analysis, tactile sensors are used for intelligent manipulation of objects. Robots can use such sensors to detect the hardness, flexibility, elasticity, roughness or shape of objects and surfaces. Robot developers can use various approaches to improve the perceptive abilities of robots. For example, using visual and tactile sensors in combination, using diverse datasets to train algorithms and other ways in which the interaction between robots and their environment becomes more organic.
There are several ways in which existing robots and future models have greater perceptibility of indoors and outdoors as compared to previous models. Indoor positioning systems using data from smart cameras and RFID tags improve the location detection capabilities of robots in indoor environments. Robots with such systems are also adept at detecting the number of persons in a room to improve interaction. In outdoor environments, computer vision tools can be used to build visual maps from Google images. Such maps can be useful for autonomous navigation and localization-related tasks. IoT, AI and robotics work in tandem to improve the perceptibility of robots in terms of sight and touch.
Sensory perception involves all senses, and not just vision and touch. To improve the audio-based perceptibility of AI and robotic tools, researchers at Carnegie Mellon University have conducted a series of tests and developed a prototype robot that could identify objects through their distinctive sounds. To develop a robot with highly-improved audio-perceptibility, the developers carried out experiments that involved robots being able to accurately identify the types of objects—such as fruits, tennis balls, metal objects and others—in a metal container by shaking it. The object identification accuracy of the robot, named “Tilt-Bot,” was 76% during such audio tests. The AI models that enabled the robot to identify objects and their material makeup were trained with thousands of datasets involving the sounds naturally created when different types of objects interacted with their surroundings. With continuous improvement, robotic applications with accurate sound, touch and sight-related perceptibility can be developed in the future.
