Intelligent wheelchair has customisable autonomy levels

The UK involvement, through Kent University, is designing a wheelchair with customisable levels of autonomy as well as devising solutions to prevent signal interference in an increasingly wireless healthcare environment.

There have been numerous previous attempts to develop autonomous wheelchairs, notably by the Massachusetts Institute of Technology (MIT) in 2008; however, they have found little real-world commercial success.

‘We want the system to be put in a new environment and learn about its surroundings as well as be able to map itself in a known environment,’ said Dr Gareth Howells of Kent.

‘If you look at the MIT stuff, it’s not that good at dealing with a large number of dynamic obstacles in its environment. If there are people rushing around in front of it, it’ll come to a stop because it can’t see a clear path — you need to be rather more assertive.’

Eschewing visual cameras due to privacy concerns, the team has opted for infrared, ultrasound and laser rangefinder sensors for navigation.

‘You can use high-level description to ask if there’s anything out there to be concerned about and if you get feedback that there is something you can try to identify what it is, where it is and where it’s going,’ Howells said.

RFID ‘tattoos’

A key aspect of the wheelchair will be integration with skin-mounted electronic radio-frequency identification (RFID) ‘tattoos’ recently developed by another team at Kent.

The RFID tattoo would communicate to the wheelchair what level of autonomy the patient requires based on his or her level of disability.

‘If there is a profile for that patient, the chair could automatically realise that, and say “OK, this patient can steer to some extent but is likely to get tired after a certain time” or “this patient is totally immobile and therefore needs total autonomy”,’ said Dr John Batchelor of Kent University. ‘If you want to rehabilitate the patient you might want to start them off with a chair that’s very autonomous but that slowly gives more and more control over time.’

Batchelor is also working with Great Ormond Street Hospital in London using RFID tattoo transfers to monitor the brain waves of epileptic children, as a more discreet method to the multitude of wires currently used. In theory, he says the RFID tattoos themselves might eventually be used to control the wheelchair with thoughts alone.

Interference issues

A key consideration of the whole project will be in preventing interference of signals in an ever more wireless and interconnected environment.

‘Until very recently, hospitals didn’t like anyone to use mobile phones anywhere — it’s been demonstrated that they can cause unusual behaviour in automated drips and ventilators,’ Batchelor said.

Batchelor heads a team that studies frequency-selective surfaces (FSS), which are specially etched metallic screens designed to allow the passage of some signals through while reflecting others away.

‘You might zone-off various areas of the hospital at various times. When the MRI device is being used, you would have a very good screen around, but when it’s turned off you might want to allow other communications in — hospitals use electronic prescriptions that can often need a Wi-Fi link.’

Batchelor is also ensuring that the wheelchair developed by Howells’ team does not create any interference in the hospital environment nor is susceptible itself.

‘If you have someone who is utterly dependent on an autonomous chair, you can’t afford to lose the signal,’ Batchelor said.

With this in mind, the wheelchair is due to undergo trials later in the year at hospitals in Kent and Lyon.

The Autonomous and Intelligent Healthcare System (SYSIASS) project is a collaboration between Kent and the Institut Supérieur d’Electronique et du Numérique (ISEN) and Laboratoire d’Automatique, Génie Informatique et Signal (LAGIS) at the Ecole Centrale de Lille, France.