MobiArm

What it does

Addresses Hong Kong's aging population challenges by providing a mobile robot assistant for daily tasks, promoting independence and reducing caregiver reliance.

Your inspiration

The concept was inspired by Hong Kong's ageing population's social isolation, mobility issues, and carer lack. Hong Kong's cramped living quarters and high-rise buildings limit geriatric movement and independence. Existing assistive technology often fail to solve these issues. Our goal was to create a mobile manipulator robot that could help elderly people with a wide range of daily tasks, promote independent living, and reduce reliance on overburdened carers while supporting Hong Kong's Smart City initiatives and sustainable solutions for an ageing society.

How it works

Multiple innovative technologies work together to power the Mobile Manipulator. The main manipulation tool is a 6-DoF robotic arm with a bespoke gripper. This arm uses the YOLO11 object detection model and a RealSense depth camera to locate things in the robot's environment. Visual information is used to plan and execute accurate grasping and manipulation motions. LiDAR-based SLAM lets the robot create a real-time map of its environment and locate itself inside it for autonomous navigation. Control is mostly done via voice command, where a powerful Large Language Model (LLM) agent processes natural language commands. This agent uses PDDL to arrange spoken orders into a plan that the robotic arm and navigation devices execute. A rising platform lets the robot dynamically modify its height, helping with sit-to-stand transitions and reaching objects at different heights.

Design process

The design process followed an iterative approach, beginning with a comprehensive analysis of the needs of elderly individuals in Hong Kong and a critical evaluation of existing assistive technologies. The initial concept centered around a mobile platform equipped with a robotic arm. The first prototype, while functional, revealed limitations in the arm's reach and stability during certain tasks. To address these shortcomings, the design was significantly modified in the second prototype, incorporating a longer frame and a U-shaped base to enhance balance during sit-to-stand transitions. The exterior was also redesigned using acrylic boards to improve safety. The gripper design underwent several iterations, evolving from a soft, deformable TPU 3-finger design to a stronger ABS 2-finger design, before ultimately settling on a CNC-machined aluminum claw with carefully selected silicone and foam padding to ensure a secure yet gentle grip on a wide variety of objects. Finally, a lifting platform was integrated to expand the robot's reach and enable it to retrieve items from the floor.

How it is different

The Mobile Manipulator stands apart from other assistive robots due to its unique combination of features and its targeted focus on the specific challenges faced by elderly individuals in Hong Kong. Unlike robots primarily designed for social interaction, such as the Care-O-bot 4, our system offers advanced manipulation capabilities. While smart wheelchairs provide enhanced mobility, they lack the ability to assist with essential daily tasks such as object retrieval or medication management. The Mobile Manipulator, in contrast, seamlessly integrates locomotion and manipulation into a single, versatile unit, making it ideally suited for the compact living spaces typical of Hong Kong. Furthermore, the intuitive voice command interface, the specialized sit-to-stand assistance feature, and the integrated fall detection system contribute to a comprehensive solution that empowers elderly individuals to maintain their independence and quality of life.

Future plans

Many significant areas will be developed to increase Mobile Manipulator usefulness. The robot's navigation algorithm for busy regions must be improved and its gripper made more flexible to safely hold more objects. Speech recognition or noise-cancellation offline may improve voice command interface accuracy and reliability. Many senior users will offer valuable feedback. We will research open-source hardware and other cost-cutting methods to minimize robot prices and boost accessibility. Finally, we want to speed up work with heart rate tracking, drug reminders, and planning algorithm optimization.

Awards

Awarded the Best Project Award in the Inter-departmental Final Year Project (FYP) 2024/25 competition, Faculty of Engineering, The Hong Kong Polytechnic University.