Future Prospects of Autonomous On-Demand Mobility

By Graham Jarvis

---

25th October 2022

Dr Shafiq Urréhman is leading all artificial intelligence (AI) and machine learning projects at CEVT AB in Sweden reveals that in his opinion there are two distinct mobility trends: logistics and customer mobility, such as robo-taxis, which he predicts will be prevalent in 2025 with limited features.

These features will include autonomous parking. Over the last five years Level 2 of automation has become widely adopted by automakers in passenger cars. Moving on to the present day, they are working on vehicle autonomy stages of Level 3 and Level 4, advanced beyond that of Tesla’s misnamed Autopilot system.

Similar trends can be seen in the development of autonomous trucks, which he thinks will become more common in 2028 in the logistics market. Although, for now, they will operate between predefined hubs as the technology is unlikely to mature enough to be given free reign.

Electric scooters, while some of their users has created some controversy owing to them being used on pavements and causing some injury to pedestrians, will also be more accepted and part of the on-demand mobility mix. “People have just started to accept the electric future with the likes of them,” he says before adding that McKinsey and Boston Consulting Group (BCG) as well as CEVT’s own internal surveys point towards this direction.

The role of data and AI

Shafiq will be speaking about these trends and the future of autonomous OnDemand Mobility at AutoTech Europe, which will be held between 15-16 November 2022 at the Leonardo Royal Hotel, Munich, Germany. He says he will mostly at talk about mostly data and its role in AI applications and services. “We will also discuss in-car applications and how other applications are being developed using data, as well as the impact of regulations such as GDPR on future data development, and there will be a panel discussion,” he adds.

Different vector considerations

Augustin K. Wegscheider, auto and mobility partner at BCG says, when considering the current autonomous on-demand mobility trends, there is a need to think about the different vectors in regard to the different models that exist. To that end he asks several questions: Is the vehicle a single occupancy, multiple occupancy, or a purpose-built vehicle for goods? Is it point-to-point or fixed route, where a shuttle operates? Point-to-point allows me to go where I want to go within a geofence, which is a robo-taxi model. Robo-taxis need to have the confidence that they can operate the software. This is why they are deployed within cities. Is the monetization model of whether it is on the per-ride basis (such as when someone gets on a bus) or the per-distance basis? This is based on the length or duration of the trip in the way that Uber charges.

Multimodal integration

“These are the landscapes of different solutions and different deployments, and all of these are being actively piloted and trialled,” he says.  From his perspective, the most interesting angle is about the extent to which on-demand mobility services complement versus compete with other solutions.

Wegscheider believes that new on-demand solutions should have a positive impact by having multimodal integration. He explains why: “You want them to be potential feeders, allowing them to take an autonomous vehicle to the train station.”  This is to increase the consumer experience and avoids any focus on treating other solutions as competition.

He adds: “In a lot of cases, on-demand mobility users are not replacing personal vehicle trips, but public transit. So, you want to integrate them as much as you can. By having an on-demand solution that takes them to a public transit system from, for example, a rural area could represent a win-win.” To encourage people to give up their own cars, mobility needs to integrate in a multimodal way. He, therefore, asks whether mobility providers and transport authorities have integrated ticketing, or whether they monetize separately.

Another option is to use Uber or Lyft platforms, or for them to create their own go-to-market app. To make on-demand mobility attractive, there need to be a high level of integration with public transport systems. At present these considerations are creating open-ended questions about the appropriate go-to-market solutions and as it’s all early in the development cycle different things are currently being trialled. He argues that it is open season because everyone is piloting.

He said: “It’s early in It’s a good time for companies to pilot and learn. You can draw parallels with on-demand, non-autonomous mobility, where there has been success with integrated booking and ticketing. This can be transferred as a “role model” to autonomous on-demand mobility.”

Total cost of ownership

Shafiq’s view of how the trends will impact on traditional mobility providers is very much focused on how mobility providers will adopt autonomous driving and robo-taxis from a total cost of ownership viewpoint. He claims that one benefit of autonomous on-demand mobility will be that customers will not have to pay higher prices for mobility from Point A to Point B.

He explains that there are four different drivers of this: “Which are paving the way for mobility to providers to go towards robo-taxis.” He says this is because the business case is very attractive and he suggests there is evidence that customers will prefer autonomous robo-taxis over other mobility forms. However, during the first phase of adoption it will be more expensive than traditional mobility. This is because the costs related to the sensors and infrastructure are relatively high at the moment.

Costs will drop

Despite this, he forecasts that the sensor, maintenance and software costs will drop. He believes the business case much will significantly improve. To this end, he explains that there is a need to factor in location, vehicle type, and the scale at which the mobility services are launched. “Last, but not least is regulation, which will play a key role in automated robo-taxis,” he suggests.

Mike Hawes, Society of Motor Manufacturers and Traders (SMMT), chief executive comments: “On-demand automated mobility services have the potential to deliver cost-effective and convenient transport. They are, however, currently limited to trials, as pilots with a safety driver present. Some of these pilots have been well received although, given their limited conditions, it has been difficult to draw definitive conclusions on how such services would integrate with existing transport options and cope with different environments.

“Regulatory reform in the UK is an essential first step before automated passenger services could be introduced on a commercial basis. Once this happens there are opportunities for public and private operators to roll out on-demand solutions, such as automated shuttles, and plug gaps for consumers where there is currently a lack of manually-driven scheduled services due to either economic or operational reasons.”

Ongoing development

For autonomous on-demand mobility to become an increasing reality, as predicted by Shafiq, there has to be ongoing work on the prediction and decision-making software. Simultaneously, work is required to change any negative perceptions of autonomous mobility. Further LiDAR and back-end operations development is essential too. “These are the major deciding factors, as well as the monetization model, for the success of autonomous on-demand mobility,” he argues.

With regard to changing consumer perceptions of autonomous shuttles, Wegscheider says there has been a lot of research on consumer acceptance of them. Unfortunately, he finds that they are falling short of giving the respondents a clear enough picture of what it would be like to ride in an autonomous vehicle. However, he claims that “when consumers are given the experience, customers’ feedback has been overwhelmingly positive”.

Addressing safety concerns

One key prerequisite for the success of autonomous on-demand mobility is safety. This isn’t just about traffic accidents. Personal safety is also a factor that can become a value proposition like that of Uber Pool. This is where people sharing a ride with others on a journey. He explains: “If you have a driver, on some level you have a neutral third-party that could mediate in cases of conflict or safety concern but if there is no driver, there is a hesitancy to share the vehicle with other riders.” This could, with regard to the Covid-19 pandemic, be about health.

These worries can be assuaged whenever a mobility provider has a purpose-built design within the vehicle to allow riders to sit separately from each other. By having such a design, mobility providers can address user expectations around safety by accounting for consumer needs and expectations around safety. After all, with passenger, users or riders, there can be no autonomous on-demand mobility, and so their views are of paramount importance to ensure its adoption and future success.