There is a high demand in rehabilitation service and daily tasks assistance due to aging. There are 2.41 billion people need Rehabilitation service today and more than 2 billion people need Assistive Technology by 2030. RRIS Robotics Lab focus on developing rehabilitation and assistive robotic technologies which can
- help the aging population to age in place while maintaining their quality of life.
- help people with disabilities to increase independence.
- help caregiver to provide better care service while reducing their physical burden.
The assistive robots developed in RRIS cover various ADLs such as transferring, commuting, walking, balancing, drinking, and eating. We also extended our target user from disabled person to health person especially elderly to help to improve their working efficiency and prevent injury by using assistive robot.
Our innovation focus on two main area: one is Novel Robotic System, which covers novel actuators, sensors, and mechanism; the other one is Human Robot Interface, which focus on using data-driven approach to achieve a safer and more intuitive way for user to control the assistive robots.
RRIS have developed several novel robotic systems which cover different ADL tasks. Our novel robotic systems consist of 2 main categories: Wearable robots and MR. X Assistant.
MR. X Assistant
Mobile Robotic Assistant (MR. X Assistant) is designed to assist users depending on their specific needs. Currently the MR. X Assistant series includes: MRTA, MRBA, MRAA and MRCA .
MRTA (Mobile Robotic Transfer Assistant) is a powered wheelchair equipped with a collaborative robotic arm that is able to carry a patient up to 90 kg via a harness system. It is designed to enable one caregiver to perform safe transfer of a max-assist patient.
MRBA (Mobile Robotic Balance Assistant) is a balance assistive robot for user who has difficulty/danger of walking due to poor balance. MRBA has 2 modes of operation: (i) Wheelchair mode that operates like a normal powered wheelchair, and (ii) Walking mode that provides sit-to-stand assistance and then act as a follow-me robot that provides balance assistance via a variable stiffness assistive robotic arm only when the user has lost his balance. The robot is capable of learning the balance ability level and gait characteristics of the user and adjust its controller to provide an optimal human-robot collaborative performance.
MRAA (Mobile Robotic Arm Assistant) is a robotic arm mounted on a mobile platform. The mobile platform can be either a standard moving platform or a robotic wheelchair also. MRAA can be used to assist users in various activities of daily living like fetching objects, feeding, dressing, cleaning e.t.c. Novel shared control algorithms will be developed to make controlling this high degree of freedom easy using intuitive but low degree of freedom interfaces like a joystick or complex inputs like speech. For use by people with upper limb disability, personalized algorithms will be developed which can adapt to user’s ability of controlling different interfaces.
MRCA (Mobile Robotic Commute Assistant) is designed to be a shared control wheelchair that can actively assist the user to manoeuvre safely and efficiently without taking away his or her control authority. MRCA is equipped with LIDAR sensors and RGB-D cameras and capable of performing the same functions of obstacle avoidance and navigation. We also developed a curb climbing module with climbing capability up to 6-inch curbs (Singapore curb is 4.9 - 5.9 inches).
Wearable robots have the characteristics of being lightweight, low powered and self-sufficient as they are designed to be a standalone device worn by the user for long periods of time. Consequently, they face more constraints than a typical robot with a large base to carry large power packs and powerful actuators. To satisfy the strict constraints, wearable robots normally provide only a percentage of the required torque for joint movement, while the user provides the remaining actuation needs. Advances in human robot interface are typically integrated into wearable robots to ensure that the assistive torque do not counteract the torque produced by the user.
3. Development and POC of Care Assistant & Rehabilitation Enabling CARE Robots
As technology becomes more user friendly, smarter, and affordable, development of assistive device is one of the most promising mobility options available today to prepare for the future aging society. The programme focuses on the development of a user-centered intelligent HRI framework for assistive robots, including a series of HRI solutions that encompasses an AI-enhanced shared control approach to enable assistive robots to provide effective assistance to help humans accomplish the targeted tasks. The programme does not intend to develop robotic systems from scratch. Instead, the plan is to work with robotic systems that are developed or being developed by research/academic institutions and companies. There may be hardware development work to enhance the sensing and actuation systems of the robots, but it is only for the purpose of developing a HRI under the proposed framework.
This programme covers three types of assistive robots:
i. Machine-on-Man – Lower Limbs & Upper Limbs, e.g., Exoskeletons
ii. Man-in-Machine, e.g. man in robotized wheelchair, vehicle, forklift, etc.
iii. Man-with-Machine, e.g. feeding robot, patient transfer robot.
and two types of people with varying physically function capability:
i. Moderate functioning, e.g. generally healthy elderly with declined strength, flexibility, stamina, vision, reaction time, etc.
ii. Low functioning, e.g. frail elderly who need assistance with most activities of daily living;
All the work packages in the programme will adopt the same intelligent HRI framework that embraces an AI-enhanced shared control approach which include:
i. predicting user intents via multiple sensing modalities,
ii. providing real-time adaptive assistance based on user’s performance, and
iii. providing safe physical human-robot collaborative control via variable stiffness mechanisms.
Human intention recognition algorithms developed based on different sensor modalities will be shared between each work package. All the work packages will work closely with RRIS to leverage on the RRIS Ability Data – a movement database of healthy people and rehabilitation patients – as a foundation to develop intelligent data-driven human machine interface.
Lastly, an extensible ROS 2 based HRI Toolbox will be developed to systematically capture the values created by all the other work packages (WP1 to WP5) within the Intelligent HRI for Assistive Robots programme so that the knowledge of know-hows, designs, algorithms, codes, etc. can be readily shared and accessed by relevant appropriate parties and re-used to facilitate and accelerate scientific research and innovation.