Can robots handle your healthcare?
Image credit: geckosystems, mit, alamy
An ageing demographic in many leading economies means that governments expect a huge increase in the demand for ‘carebots’ in health and social care. But is the technology up to the task?
As life expectancy increases, particularly in wealthier countries, a shortage of carers is becoming alarmingly evident. If the problem of dealing with an ageing population isn’t addressed soon, there will be massive consequences for elderly and sick people and those who care for them, as well as for healthcare systems worldwide. Japan is the first nation to be hit by what has been somewhat tastelessly called the ‘silver tsunami’, with more than a quarter of its population aged over 65, and the proportion in many European countries is topping 15 per cent.
Faced with this challenge, Japan is building on its long-established expertise in industrial robotics to spearhead a unique solution - mobile service robots for healthcare. In an effort to put the patient at the centre and front of this programme, the Japanese government has constructed 10 development centres where elderly ‘care receivers’ or patients will share ideas with development support coordinators and robotics engineers to help make the robots more patient-friendly and efficient.
There are number of similar initiatives around the world, exploring both the potential benefits and risks of these ‘carebots’ to the world of healthcare.
Introducing robots into the healthcare system might seem to be an ideal solution for an ageing population, but it does bring its own challenges, including issues around safety. Martin Spencer, CEO of Gecko Systems, notes that in order for any companion robot to be practical and useful enough for family care, it must be ‘like a three-legged milk stool’ - in other words, safe for routine use.
Spencer also believes that for a mobile robot to move in close proximity to humans, it must have quick enough reflexes to navigate around unexpected items within the environment, have verbal interaction capabilities that include an awareness of the date and time, have the ability to track down people either nearby or remotely and be able to interact with people by audio and video. He says Gecko’s Carebot is one such robot. It has an inbuilt tracking system that locates patients using a combination of machine vision, infrared and sonar sensors. This also allows it to navigate the environment safely.
To remain close to the patient, the Carebot’s infrared sensor scans for and detects the body heat of a person. The robot’s vision and memory systems are also sophisticated enough to allow them to identify and remember the clothing pattern that a person is wearing. To ensure the robot does not trip anyone up, a sonar range-finder comes into play to maintain a safe distance. “A high level of situational awareness is achieved through blending several sensor systems,” says Spencer.
For example, if a mobile service robot is going to the kitchen to remind the patient to take pills and there is a chair in its path, it can make the navigational alterations to avoid a collision, get back onto its original course and deliver the message.
It accomplishes this through artificial intelligence (AI) software engines, which combine the specialised knowledge of an expert with input from several sensors that pick up information from the environment.
In order to compensate for uneven surfaces and maintain its target destination, the robot’s navigation software adjusts the power sent to each wheel of the device. This navigation system is made up of various cameras, detectors, range finders and a number of other systems that communicate by a Wi-Fi network router.
This technology allows a mobile service robot to check on a person if they start moving around at night and to notify human carers. If they are not on the premises the robot communicates remotely with a designated person who can then assess from afar whether there is an emergency that requires a visit.
Additionally, the Carebot can be configured to call emergency services immediately via the internet.
Care and patience
Monitoring patients via varying sensors and even reminding patients to take their medication is perhaps something that has been a long time coming, given rapid advances in technology. However, what about robots that can help patients physically, just like a human caregiver?
The Robear, a plastic and metal robot with a cute polar-bear-cub face, provides patient-transfer functions for people who cannot get out of bed unaided. It was developed by Japanese research institute Riken in collaboration with the Sumitomo Riko Company.
The robot can lift patients in and out of bed, placing them into a wheelchair or assisting them if they are unsteady on their feet. This provides relief for carers who often have strained backs from lifting patients on a regular basis.
Robear has been given a smaller base than earlier versions to reduce the overall weight of the system, but with legs that can be extended to make it more stable when lifting a patient and retracted for manoeuvring in confined spaces such as doorways.
The reason this human-like robot does not have a human appearance, says Dr Toshiharu Mukai, research leader for the Robear project, is “because imperfect human-like appearances give users a strange feeling - the so-called ‘uncanny valley’”.
Mukai says there is still a long way to go in refining Robear before it can be deployed in nursing homes. He explains: “Robear was built for research purposes and there are still large gaps between research experiments and practical uses. For example, Robear is too complicated and maintenance is too difficult for practical use.”
Helping children manage pain
Robots can excel humans at providing care in terms of patience under emotionally difficult situations. While a human may lose patience towards the end of a long shift with yet another screaming child refusing medication, a robot can retain the same soothing tone in order to persuade the child of the benefits of the medication.
Medi (Medicine and Engineering Design Intelligence) is a 60cm-tall humanoid robot, produced by RxRobots, that is used in several hospitals and dental offices in Canada and the USA. Dr Tanya Beran, founder of RxRobots, explains: “The social behaviours that express emotions are key to Medi’s success.” Movements of the arms and hands are associated with what Medi says. Also, the LED lights in the eyes can suggest emotions of shyness, surprise, sadness and sleepiness. The lights, arm and head movements and eye colours work together to recreate expressions which humans are familiar with and can relate to.
Beran claims that Medi has been shown to reduce children’s pain perception by as much as 50 per cent in clinical trials. Furthermore, the integration of psychology-based software allows the robot to interact appropriately with children in medical and dental settings. “It has the ability to teach the patient pain management strategies in order to relax them before, during and after a procedure, play music and even games,” she adds.
There have been some design changes to improve usability since Medi was first introduced. Beran says: “To operate the robot, people had to touch the sensor on the head and scroll through the list of applications. This is cumbersome. Instead, we provide the solution of a tablet that lists all the behaviours for the robot (like apps on a cell phone) and the user can simply see the behaviour and click ‘play’ to make the robot carry out that behaviour. The tablet communicates with the robot via a router.”
Asked whether Medi has been found to provide better interaction with children than a robot that more closely resembles a human, Beran was of the opinion that “if the robot’s features closely resemble a person, this creates a freaky feeling, especially for children. We avoid this problem by providing an adorable, endearing, polite, and friendly robot that’s just the right size - so even children as young as four years of age are not frightened. Being gender-neutral also sparks children’s imagination about whether the robot is a boy or girl and what types of personality characteristics he/she may have. In this way, the robot becomes a creative platform for interaction between a healthcare worker and the child and family.”
While some robot designers have chosen to stay away from a human resemblance, ‘Nadine’ is an extremely human-like mobile service robot. Nadine was developed by Nadia Thalmann, a professor at the Nanyang Technological University’s School of Computer Science in Singapore, who says: “Nadine is very much like a human, but she is somewhere between a machine and a human. She has the appearance of a human, behaves as such, but in reality she is an embodied computer that has all the capacities of a computer.”
Her most useful functions in healthcare, according to Thalmann, are to “monitor the patient, talk to the patient, be able to read stories and communicate with and report to family members.”
Thalmann goes on to explain that Nadine uses Google Chatbox for recognising what is said. Then the words are analysed in relation to processes stored in the robot’s memory and a decision is made as to which of the robot’s multiple possible modes of interaction should be used to respond, in terms of movement, speech synthesis, ‘facial deformation’ and gestures. However, Thalmann admits, “Nadine is not always able to understand what we are talking about. She also makes links with facts she has in her database so sometimes she answers in an incorrect way.”
At the latest International Conference on Robotics and Automation in May 2016 in Stockholm, Sweden, Massachusetts Institute of Technology’s (MIT) Professor Daniela Rus, together with Shuhei Miyashita, formerly of MIT and now a lecturer at the University of York, spoke of what they call an origami robot - a robot contained within a one-centimetre pill capable of being swallowed by a patient.
Once inside the body, the robot unfolds and can perform different functions such as removing a small button-type battery from the stomach. According to the Washington-based National Capital Poison Centre, around 3,500 people of all ages have swallowed button batteries in the US alone. The electric charge within such batteries can react with saliva and stomach acids, resulting in burns and even death. A number of tests have shown the origami robot to be successful in removing a button battery from an artificial stomach. To do this, the robot’s movement is controlled with a magnetic field. In tests, it took around five minutes to locate the battery lodged in the artificial stomach. So far, the device has not been tested on humans, but for tests on pigs, the researchers plan to use sonar and magnetic resonance imaging (MRI) to help them locate objects that need to be removed.
The need for human creativity
So are robots the future of healthcare? At least for routine tasks, it looks as if they are likely to become common in healthcare settings. As machine learning expert Anthony Goldbloom recently argued, robots are well-suited to absorbing a large number of examples or books on a subject and applying that knowledge to a situation. Often with more precision than a human, they can navigate obstacles, perform repetitive tasks to a specification, identify and remember patterns and be designed to go places where humans cannot.
Although their appearances and behaviours may eventually become indistinguishable from those of humans, some believe they will always lack the human ability to think creatively and ‘out of the box.’ In the healthcare context, this raises the question of whether robots will ever be able to truly replace caregivers - will they be able to empathise with a patient or be able to identify environmental causes for disease? Such questions remain to be answered, but the scope for mobile service robots is certainly an area to keep an eye on. *
What are the future prospects for robots in healthcare?
AI for wheelchairs
Gecko Systems is taking the sensor-fused navigation system used within the Carebot and adding it to wheelchairs to help people avoid collisions when navigating crowded spaces, especially elderly people, young children and people with neurological diseases such as multiple sclerosis. The SafePath system takes over and navigates around obstacles, keeping the person in the wheelchair safe, while at the same time giving them a measure of independence. The CareChair, as it will be known, also has sensors to measure heart rate and blood pressure, so that nursing staff can be alerted of any changes requiring immediate attention.
A speech-making robot
Huge strides have been made in linking a robot’s voice to facial expressions. Dr Tania Beran, founder of RxRobotics, explains: “Our Medi robot was originally conceived as a coach and distraction for pain management, but people have extended its use to physiotherapy, educating families about health, providing mental health support. It also is a speaker for special events – Medi thanks people for their generous donations, welcomes them to fundraising events and tells jokes at dinner functions. The possibilities are almost endless.”
However, Martin Spencer of Gecko Systems warns that with medical advances based on robotic innovation, there is a flip side - sometimes dangerous AI behaviours are possible. For example, what if a medical robot were to decide that it would rather hurt than help? Although such dangers can also arise with human carers, AI-enabled robots might, one day, think and mobilise into groups in ways that humans cannot.