Adaptive Virtual Reality for Autism: A Heart Rate Driven Sensory Regulation System Using Large Language Models

As my research has developed, I have become increasingly interested in the idea of personalised regulation. Earlier research into airports showed that sensory rooms, quiet spaces and other accessibility interventions can help, but they still tend to provide one shared environment for people whose sensory needs can be very different. This led me to look more closely at adaptive virtual reality and whether VR could do more than simply simulate an airport or prepare somebody for travel.

I found a systematic review examining adaptive VR systems for autistic people. What interested me most was the emphasis on autism being highly individual, meaning that one fixed intervention is unlikely to suit everyone. Adaptive VR offers a different approach because the environment itself can potentially change depending on the person using it.

The review included 10 studies involving 129 participants, 76% of whom were autistic. Across the studies, 70% used level switching, where the VR experience changed according to the participant’s progress, and 90% used adaptive feedback strategies. Some systems adapted using explicit information such as task performance, while others used implicit signals including movement, eye gaze, speech and physiological responses. This suggests that adaptive VR can potentially respond to how a person is behaving or reacting, rather than requiring them to manually communicate every time something needs to change.

This connects strongly to my own research because one of the recurring issues has been the difficulty of creating one sensory environment that works for every autistic traveller. One participant in my primary research commented that, “The sunflower lounges are often still quite open and loud and not tailored to autistic people/sensory needs.” This made me question whether shared sensory spaces provide enough individual control. It also reminded me of many experiences ive had with my sister, one in particular was when we tried to go to an autism friendly screening at cinnieworld, the other children with autism at a differant level or need to my sister made so much noise that my sister found it harder to regulate than in a normal cinema room. There are different levels of need. Adaptive VR introduces a different possibility: rather than giving everyone the same ‘calming’ environment, the experience could potentially respond differently depending on the individual using it. Could we design one solution, to differant levels of need?

This is particularly relevant to air travel because airports are inherently difficult to control. Delays, crowds, queues, announcements, gate changes and unexpected changes cannot all be removed through design. If the external environment cannot always be changed, perhaps there is value in exploring whether the individual’s immediate sensory experience could become more controllable.

At this stage, I do not want to assume that VR is the final answer. The review does not prove that adaptive VR would work for independent autistic adults travelling through airports, and the studies were relatively small and often focused on younger participants. For me, the value of this research is more in the principle it introduces: an environment does not have to remain fixed. It can potentially respond to the person using it.

This has started to shift the way I am thinking about accessibility. Instead of asking how one airport space can be designed to suit every autistic traveller, I am beginning to ask whether accessibility could become more personalised and responsive. A question I want to carry forward is:

Could the principles of adaptive VR be transferred into an airport environment to support personalised regulation during the live journey?

Adaptive Virtual Reality for Autism: A Heart Rate Driven Sensory Regulation System Using Large Language Models — IEEE, 2026. This is the one where the VR can alter things like