Today, there are over 25 million Americans with travel-limiting disabilities, and of this group, only a fifth are able to work due to the severity of their conditions [1]. While the cause of disability may vary, those who have lost movement ability share several key needs. First, patients need tools that provide the option for therapeutic exercise so they can recover lost function, if possible. Second, tools need to provide a motivational component to keep patients engaged in the recovery process. Third, tools must provide a way for patients to move about life as they recover [2]. Considering that most disabilities also come at a high lifetime cost in addition to limiting the ability to work, as is the case with traumatic spinal cord injury (TSCI) where costs can range in the millions for a single patient, tools must also be affordable to be effective [3-6].

Past attempts to solve these challenges have come in the forms of design modifications to the wheelchair to provide therapy and even more technical creations like exoskeletons which assist paralyzed users to walk [2]. Autonomous vehicles pose another attractive answer that can provide even greater independence than aforementioned solutions. Given that automobile accidents are an overwhelming cause of some of the most devastating disabilities like TSCI, autonomous vehicles could even serve as a preventative measure. Considering that 78% of patients with TSCI are young men under the age of 30, an autonomous vehicle truly designed for disability in mind could mean the difference of a lifetime of experiences [3-6].

In this article, I review statistics on disability transportation and discuss the special considerations that must be taken to design autonomous vehicles for accessibility. Next, I discuss current trends and challenges in autonomous mobility, and how location technology specialist TomTom is working to move autonomous driving forward with its high-definition (HD) maps. Finally, I share my recommendations to the industry for next steps.


Current state of disability transportation

Private vehicle ownership remains the chief mode of travel for Americans, with non-disabled individuals using personal vehicles slightly more (83.9% of the time) versus individuals with disabilities (74.8%) [1]. Considering that income is a primary determinant of vehicle ownership, it is unsurprising that 12.2% of disabled workers live in zero-vehicle homes, a near three times increase over the 3.9% statistic reported for non-disabled individuals.

Vehicle ownership is further complicated as most private vehicles require aftermarket modifications to make them accessible to users, especially those with wheelchairs. In his New York Times article, Henry Claypool, a policy director for the Community Living Policy Center at the University of California, San Francisco and wheelchair user, detailed how these modifications cost him an additional $25,000, which highlights these challenges of accessibility and cost [7].

To combat high costs, those with disabilities employ various strategies to get around. Options like bus lines and paratransit are more frequently used in this population than in unimpaired individuals, making up 4.3% of the mode share of travel for disabled people, versus only 2.7% of travel for the unimpaired. Local transit can be problematic, however, due to delays and routing times [1]. Ridesharing services offer another possibility, but higher fares in large cities can make services prohibitive.

Another factor is the time-cost of being disabled. One study showed that 45% of TSCI patients require assistance to transfer in and out of their wheelchair [8]. Considering, then, the amount of transfers that must take place for a disabled individual to even exit their home via car, it is unsurprising that nearly 3.6% of the 25 million people with travel limiting disabilities simply stop driving or leaving home [1]. This then further limits employment opportunities, underlining a negative vicious cycle and emphasizing the need for autonomous vehicles for disability.


Trends, challenges, and recommendations in autonomous mobility for disability

Though it is widely recognized by the industry that people with disabilities could benefit from autonomous vehicles, more special knowledge is required to design truly helpful cars. As evidenced from the language used on websites for self-driving car companies, disability is presented as a monolith. For example, there is little content available on these websites describing what specific disability these companies’ cars are targeted to help. These companies almost always show self-driving cars that are standard vehicles, not modified for wheelchair users’ accessibility.

Of similar frequency, these companies post jobs with titles like “rehabilitation engineer” or “disability advocate”—positions that could go a long way in providing a voice of a customer truly in need. Given the prohibitive cost of designing an autonomous vehicle, it is perhaps unsurprising that these roles have not yet been created. Key technologies needed to make autonomous vehicles operable and safe, such as LiDAR sensors, can cost up to $70,000 per sensor.

In itself, the tricky combination between building an affordable product and a safe one is a constant challenge to the industry, and the clear potential for autonomous vehicles to help people with disabilities is sometimes overshadowed by the accidents that have occurred. As sensor costs decrease and progress is made by self-driving car companies, other concerns specific to disability will arise and must be addressed to make truly safe vehicles. Face tracking, a technology currently used to monitor driver awareness in case human intervention is needed, may need to be specialized to the needs of people with disabilities. Car makers must also ask how extra safety can be provided given that some disabled users may never be able to take the wheel at all [9].

A possible solution to the sensor debate may come in the form of optimizing resources so these safety-critical algorithms can occur efficiently. At TomTom, we provide benefit to self-driving car makers via our HD maps, which provide high-definition representations of the world needed for the localization of the car. In addition to providing the data to route the car to reduce travel time, the map data are compressed so more computing power can be spent on sensors processing other necessary safety functions. Thus, by compressing the data, TomTom may be able to provide another solution in the core challenge of autonomy, balancing cost and safety.



For the millions with disabilities, autonomous vehicles hold a promise of new independence and even potential prevention of injury, but challenges remain. The high costs of self-driving vehicles limit accessibility, and safety remains a chief concern. Despite the overwhelming need for disabled user-centric design, autonomous vehicles have largely ignored these requirements in their initial production stage. To address these needs, the industry must build teams that can accommodate a larger perspective, by including rehabilitation engineers, physical and occupational therapists, as well as include input from users themselves. Reducing cost via approaches such as data compression, as seen in TomTom’s HD maps, so resources can be focused on safety, is another potential solution—all of which can lead to a brighter road ahead for autonomous vehicles and disability.



[1] Brumbaugh, Stephen. "Travel Patterns of American Adults with Disabilities." Issue Brief (2018).

[2] Sarigul-Klijn, Yasemin. (2018). Gait Rehab Adaptive Machine: Design of GRAM, a Walking Linkage Powered Wheelchair for Lower Body Therapy and Assistance. V001T03A003. 10.1115/DMD2018-6816.

[3] Ahuja, Christopher S., et al. "Traumatic spinal cord injury." Nature Reviews Disease Primers 3 (2017): 17018.





[8] Gerhart, Kenneth A., et al. "Long-term spinal cord injury: functional changes over time." Archives of physical medicine and rehabilitation 74.10 (1993): 1030-1034.

[9] Bradshaw-Martin H., & Easton C., "Autonomous or 'driverless' cars and disability: a legal and ethical analysis", (2014) 20(3) Web JCLI.