New Dexterity, a research group located within the University of Auckland’s Faculty of Engineering is changing the game for robotics in New Zealand and across the world. Dexterity means having the skills to perform complex tasks that need fine motor control and movements, especially with hands. Our hands play a crucial role in being able to achieve small goals and complete basic tasks without help from others. New Dexterity is aiming to make both humans and robots more dexterous.
New Dexterity is directed by University of Auckland Senior Lecturer Dr. Minas Liarokapis and is made up of 11 core team members and several other students.
The goal of New Dexterity is to develop and innovate within the robotics space, specialising in replicating, or enhancing human dexterity. The group has researched and developed robotics and technology in many areas including prosthetic arms, grippers and hands, wearable robotic exoskeletons, and teleoperation and telemanipulation.
From left to right: Geng Gao, Mojtaba Shahmohammadi, Gal Gorjup, Dr. Minas Liarokapis, Lucas Gerez, Anany Dwivedi, Che-ming (Ben) Chang
Their latest innovation is a human-like robot called ARoA. In English, ARoA stands for Autonomous Robotic Assistant and in Te Reo Māori Aroā means perception and comprehension. Dr. Liarokapis says ARoA aims to be an autonomous intelligent robotic agent that can execute many activities that are a part of daily living with increased dexterity.
ARoA currently weighs 285kg and has been in development since 2019. She (as per Dr. Liarokapis's introduction) includes many elements New Dexterity has created through rigorous testing. Che-ming Chang (Ben) is a PhD student who worked on the anthropomorphic arm and hand systems as well as ARoA’s hardware and software alongside other PhD students including Gal Gorjup, who architects the perception and operating software, Geng Gao, who develops novel gripper designs for complex grasping, Lucas Gerez who worked on the multi-modal torso, Anany Dwivedi who worked on the control and perception part, Jayden Chapman who worked on the robotic head, and finally Dr. Liarokapis.
While still in the development phase, ARoA can presently complete cleaning tasks, tidying up tasks, disinfection tasks, and grasping and manipulation. Dr. Liarokapis says “The end goal for ARoA is to be able to accomplish as many activities of daily living as possible and to be an intelligent robotic assistant for people that need her (e.g. elderly, people with limited mobility, etc)”.
Dr. Liarokapis says ARoA has been intentionally designed to look friendly and approachable. Alongside PhD student Jayden Chapman, they produced a cartoon-inspired head to reduce any scariness of the robot. “Due to the Uncanny Valley of anthropomorphism. A more human-like face on a robot would look creepy.”
The Uncanny Valley is a concept that was introduced in the 1970s by Masahiro Mori, it means that human representation is good and appealing in robots until a certain point (Uncanny Valley) where it then becomes too real for people to happily comprehend. Human anatomy and shape have been used for the rest of the ARoA’s body e.g. head, spine, shoulder and elbow joints, hands.
The future for ARoA is bright. She is an innovative piece of technology that aims to assist in the service industry such as hotels and hospitals. The next step is to test ARoA in these environments. After a testing period, the team at New Dexterity can evaluate the data and make improvements where needed to make ARoA fully fit for purpose.
What helps ARoA perform her tasks effectively are the hand-like grippers designed and created by New Dexterity.
CONTROLLING ROBOTS WITH OUR MUSCLES AND OUR MINDS?
PhD student Anany Dwivedi is completing a PhD in muscle machine interfaces at Auckland University and has since built EMG based interfaces that can control the robot in an intuitive manner.
EMG (electromyography) is used to detect muscle movement. Sensors are attached to the skin of a human's arm and they pick up muscle movements the robot is then able to mimic. This technology is currently being used in ARoA.
The EEG (Electroencephalography) cap is placed on a human controller's head. Electrical signals travel from the brain through a conductive gel to the sensors. From here the aim is to decode the signals into an intention for movement so the robot can be controlled. The end goal of this technology is to allow a human to control robotic helpers using just their thoughts. This technology is still in development.
Alongside ARoA, many other innovations in the robotics world are happening within the New Dexterity research group. PhD student, Lucas Gerez is developing exoskeleton gloves. Gerez has developed three prototype gloves that aim to support people with hand injuries, disabilities, as well as reducing fatigue and risk on worksites.
The first prototype glove has an inflatable exterior and a second thumb (like a koala) that allows the grip to remain rigid without causing the controller's hand to become tired.
The second version of the exoskeleton gloves has a camera in the palm that uses artificial intelligence (AI) to identify what the glove wearer is about to pick up. The AI inflates the octopus-like grip pads to the correct shape and grip so as to pick up that object. While these were inspired by octopus suckers, instead of creating a vacuum action to suck, they inflate to provide the best precision grips.
The third glove Gerez has designed is powered by force generated by the user. This particular device allows the user to experience enhanced strength of the hand. By using force and energy provided from the user’s fully functioning shoulder the force travels down a wire-like structure and applies the force to the user's weaker hand.
DRONES THAT CAN GO ANYWHERE
New Dexterity PhD student Joao Buzzatto has been working hard to develop new UAV (unmanned air vehicles) technologies, also known as multi-terrain drones. His creation called the New Dexterity Coaxial OmniRotor UAV can apply force continuously in any direction. It can even fly upside down. It is also capable of propelling itself along the ground using the
wheels and rotors. Joao hopes this drone technology will be used to reduce the risk humans face in inspections and maintenance of hazardous construction areas as well as surveying natural disaster and recovery areas. Navin Perera, a Masters researcher is helping Buzzatto with design and fabrication. The first prototype was put together in March 2020.
Joao Buzzato (right) with the second iteration of the Coaxial OmniRotor completing flight experiments conducted in collaboration with Pedro Mendes (left) in the UAV flight testing lab.
New Dexterity will keep innovating, creating, and developing cutting edge robotic technology that aims to help the world be a more convenient and accessible place for everyone. When asked where they see ARoA in one year, Dr. Liarokapis said, “We will have a fully developed second version of the platform that will be able to execute a wide range of everyday life tasks.”
Thanks to the work of New Dexterity and others in the robotics field we can expect to see robot assistants and enhancements becoming a part of our lives, helping us perform tasks with ease and skill and reducing the need for humans to complete dangerous, repetitive, uninteresting and, well, 'robotic' tasks.
Story by Makayla Wallace-Tidd. Photography by Richard Ng
CITE THIS ARTICLE:
Wallace-Tidd, Makayla. Human-like Robots or Robot-like Humans? New Dexterity is Bringing us Closer. MOTAT Museum of Transport and Technology. First published: 07 November 2020. URL www.motat.nz/interactive/new-dexterity