Team Member, S-HERO Co-oP project in Partnership with [WeGo Robotics]
Role: Hardware Design
Received Excellence Prize (SKKU S-HERO Corporate Problem Solving Program)
Key skills: 3D CAD / Hardware Mechanism Design
Abstract
With the advancement of medical technology, the average lifespan has increased, and coupled with declining birth rates, many developed countries including ours are experiencing an accelerated aging population. Moreover, there is an ongoing shortage of nursing care workers. Lastly, the personal and societal economic burden of elderly care is projected to rise significantly. However, in such a scenario, if logistic robots could replace some of the duties of nursing care workers, tasks like transporting heavy items including bedding could be eased, and many errands could be substituted, saving a considerable amount of labor. This substitution could potentially alleviate the workload of nursing care workers, improve their working conditions, and reduce the social and economic burden of elderly care. We have proposed a robot to assist nursing care workers in nursing homes by transporting various items. With a two-tiered storage space, the robot can carry multiple items at once, and by utilizing SLAM (Simultaneous Localization and Mapping) technology, it aims to achieve indoor autonomous navigation. This robotic assistance in handling logistics within nursing care facilities not only augments the efficiency of care but also paves the way towards integrating modern technological solutions in addressing the challenges posed by an aging society.
Robot Hardware
The robot hardware is divided into two main sections: the goods storage area (upper part) and the mobile robot base. For the mobile robot base, we adopted Wego Robotics' ERP42, and designed the goods storage area with two spacious levels. To ensure safe transfer of items out of the robot upon arrival at the destination, an elevator mechanism utilizing a Scissor & Tilt design was incorporated within the central part of the goods storage area. When the elevator reaches the desired level, a conveyor belt situated below the goods activates to move the items onto the elevator. Subsequently, the elevator lowers the items, and utilizing its tilt feature, allows the goods to slide out of the robot smoothly. This design ensures a systematic and safe process for transporting and delivering items, making the robot a reliable assistant in logistics within a nursing care environment.
Scissor & Tilt Mechanism
The use of a scissor mechanism in the elevator was chosen to facilitate the stable transfer of goods within a confined space. By doing so, the length of the scissor lift can be adjusted by moving the linear actuator located at the bottom part. Moreover, to ensure the safe dispatch of items outside the robot, a tilt feature was incorporated. This tilt functionality was designed to be operated by an additional linear actuator, enabling a controlled tilt for a smooth and safe transfer of goods out of the robot. Through these mechanisms, the robot is equipped to handle goods efficiently and safely, making it a practical tool for aiding in logistics within care or other facility environments.
Mapping and Localization
Initially, a map of the nursing home where the robot will operate is required. However, due to COVID-19 restrictions, direct mapping within the nursing home proved to be a challenge. Consequently, an area with a similar structure to the nursing home was selected for mapping. Specifically, the corridor and some rooms of the basement first floor of Engineering Building 1 were mapped as a stand-in for the nursing home. For the mapping process, the ROS (Robot Operating System) package rtabmap was utilized.
Localization and Navigation
Since the robot is designed for indoor navigation, collecting GPS data is not feasible. However, leveraging the pre-existing knowledge of the travel area and path, the robot's localization was executed using the pre-mapped data of the area. The flow chart of the robot navigation algorithm I implemented and the overall system architecture can be seen in the figure below. Initially, within the `mission.py` code that manages the robot's navigation endpoints, various scenarios were planned in advance and stored in a dictionary format, delineating the robot's paths for each mission. From this setup, upon receiving a mission key, the robot sets its destination and the path to achieve it.
The primary code used during navigation is `pathfinder.py`, where the robot is programmed to follow the designated target point. When the platform steers either right or left, local replanning is performed considering the changing position of the destination based on the turning radius. Also, during this process, if static or dynamic obstacles are detected, an avoidance radius is set considering the size of the robot, subsequently generating an avoidance path and performing local path replanning. The calculated values for speed, steering, and braking during these processes are transmitted to the platform through serial communication in `control.py`. When transmitting steering values, consideration is given to the platform's wheels' maximum possible steering angle, and if a derived steering value exceeds this maximum angle, it is limited to the maximum steering angle before being sent to the platform. This design ensures that the robot navigates safely and efficiently within the indoor environment, adapting to obstacles and executing its missions as planned.