An Autonomous Humanoid Robot Designed to Assist a Human with a Gesture Recognition System

Abstract

This paper presents the design of an autonomous humanoid robot designed to optimize and enrich customer service in showrooms, e.g., electronic equipment, mobile network operators, and generally in stores with various articles. The proposed humanoid robot design is distinguished by two key components: a sensor-equipped mobile platform with drives and a body featuring a head outfitted with a touch tablet and an RGBD camera. The control system enables autonomous navigation in both known and uncharted environments, with a special focus on diverse, crowded, and cluttered spaces. To enhance its adaptability, this robot is not only fitted with LIDAR sensors but also cliff and ultrasonic sensors. While the interactive ability with humans is an expected functionality, this paper brings forth certain distinct innovations in humanoid robot design for customer service. One of these unique aspects includes the robot’s ability to physically alter its configuration, such as rotating its head and adjusting the height of its torso to maintain line-of-sight with the customer. This capability signifies a novel degree of spatial responsiveness that exceeds static interaction. Moreover, the proposed robot is equipped with a user-friendly gesture recognition system, uniquely designed to detect and recognize simple human hand gestures. This attribute paves the way for understanding simple commands such as requests for assistance. Upon recognizing a request, the robot tailors its services by following the person around the showroom, effectively assisting and answering customer queries or displaying requisite information on its screen. This active assistance model, specifically tailored for human interaction, showcases the robot’s unique capability to respond proactively and dynamically to human inputs.

Time division multiplexing based multi-spectral semantic camera for LiDAR applications

“…The 2D convolutional filter shares weights in the x and y dimensions to preserve spatial information and features while maintaining associations between neighboring pixels 41 , 42 . The excellent object recognition performance of CNNs makes them highly suitable to provide object recognition capabilities for autonomous vehicles, which are crucial for detecting potential obstacles such as pedestrians, other vehicles, and bad road conditions using a camera sensor 1 , 2 . However, a conventional camera image lacks the distinction between objects of different materials but of similar shape and color that frequently appear while driving.…”

“…Autonomous recognition systems have made remarkable progress in recent years and have various applications including robotics 1 – 4 and driving systems 5 – 8 . Autonomous driving solutions have employed camera 9 – 14 or light detection and ranging (LiDAR) systems 15 – 19 to acquire the surrounding environment information during driving, which provides crucial feedback for navigation and safety strategies such as routing and collision avoidance.…”

A Bio-Inspired Chaos Sensor Model Based on the Perceptron Neural Network: Machine Learning Concept and Application for Computational Neuro-Science

“…The time series is input to the perceptron, and the entropy value is calculated at the output. Such sensor can be applied not only in the field of computational neuroscience, but also to create humanoid robots [ 34 ] and animal robots [ 35 , 36 ], including devices with limited resources [ 37 ], as well as to improve neuroadaptive learning algorithms for robust control in constrained nonlinear systems [ 38 , 39 ]. For example, in biosimilar models where information is distributed as a spike train, changes in the packet frequency are often studied [ 40 , 41 ].…”