By: Eric Paul Dennis, PE – epdennis@crcmich.org

In a Nutshell:

• The State of Michigan and many local governments have been preparing for autonomous vehicles (driverless cars) to hit public roads, diverting resources and attention away from other matters.
• Recent setbacks in the industry show that any transition to autonomous vehicles will be slower and less dramatic than previously expected.
• Michigan transportation planners should refocus efforts on proven solutions to today’s problems.

Michigan prides itself as the home of the automotive industry and a leader in transportation technology. This image is important for many Michiganders, and policymakers are keen to stay ahead of emerging trends in automotive and transportation. Many of Michigan’s road agencies are anxiously preparing and planning for a mobility revolution brought about by autonomous vehicles (AVs). There is an expectation that driverless/self-driving cars will solve many of our transportation problems by making travel safe, efficient, and equitable.

Such expectations were premature if not misguided. It will likely be many years before AVs become a meaningful component of the transportation system. Policymakers should refocus on known solutions to address existing problems.

Popping the AV Hype Bubble

Innovators have been working for decades to make driverless cars a reality. The recent investment bubble kicked-off in 2012, when Google released a video demonstration of a blind man using a self-driving car to run errands. Many had believed such technology was decades away if possible at all. The initial successes of Google’s self-driving car program preceded a flurry of research, development, and investment activity across multiple industries. New startups, established technology companies, and automakers joined a race to be the first to put self-driving cars on the roads. 

As real-world difficulties have become apparent, timelines have been pushed back indefinitely and investments have nearly dried up.

Now, a decade on, it is widely estimated that over $100 billion has been invested in the effort. However, as real-world difficulties have become apparent, timelines have been pushed back indefinitely and investments have nearly dried up. The future just ain’t what it used to be.

Many of the boldest and most attention-grabbing efforts have been lost to consolidation or bankruptcy. In October 2022, Argo.AI became the most recent self-driving car company to close down.

Argo’s demise was a big deal. Many industry-watchers considered Argo to be an industry leader. Both Ford and Volkswagen had invested billions of dollars in the effort. Argo continued to announce milestones and forward-looking plans right up until Ford announced that it “made a strategic decision to shift its capital spending” and would be writing down $2.7 billion on the investment. Ford President and CEO Jim Farley explained the move by saying that, “fully autonomous vehicles at scale are a long way off.”

AVs and CAVs

The State of Michigan government has invested tens of millions of dollars and initiated multiple efforts to support the development and deployment of connected and autonomous vehicle technology. These efforts have evolved over the years and are now integrated into the Michigan Connected and Automated Vehicle (CAV) program. 

The term “CAV” is sometimes used as a shorthand label to refer to a range of vehicles with connected or automated technology. The Michigan Department of Transportation (MDOT) has adopted a formal definition for their latest CAV Strategic Plan (2021):

“Connected and automated vehicles (CAV) employ both automated and connected technologies. These two technologies work together cooperatively to further enhance the safety benefits offered by each, as follows: While automated vehicles are expected to improve vehicle safety by limiting the impact of human error, connectivity enables additional safety benefits, as vehicles can then gain context beyond what a regular driver would know or have the ability to perceive visually. Similarly, while connectivity can enable alerts and warnings to a driver-operated vehicle, deploying these messages on an automated vehicle can streamline the links between information, decision making, and action.”

In other words, Michigan policy defines a CAV as a vehicle whose control systems are integrated with its communication systems. For example, a CAV might brake if it gets a message from a connected signal, or multiple CAVs might coordinate speeds and positions to form a platoon. 

In theory, a “connected automated vehicle” leverages advantages of both connected and automated vehicle technologies.

Image source: USDOT via Oregon DOT

As previously noted, autonomous vehicles are no longer viewed as a near-term possibility. Similarly, connected vehicles that communicate directly with nearby vehicles and infrastructure (as described in the image above) are not likely to be deployed anytime soon. The communication technology being considered for driverless cars is a different system than the cellular connection that many of us have in our cars. This technology remains in research phase and may never be commercially adopted. Combining these two technologies is an especially unlikely future.

CAV Corridors

Despite the recent collapse of the AV hype bubble, and negligible commercial interest in CAVs, the State of Michigan continues to pursue policies intended to leverage these technologies to reimagine transportation systems and drive economic development. The most ambitious of these projects is the Connected and Automated Vehicle Corridor (CAV-C) effort. MDOT has selected an approximately 25-mile segment of Interstate 94 in southeast Michigan (between Detroit and Ann Arbor) for the initial deployment of the CAV corridor. It is envisioned that the dedicated corridor (pictured below) would provide an infrastructure solution that allows for a mix of connected and autonomous vehicles, traditional transit vehicles, shared mobility, and freight and personal vehicles.

CAV-Corridor dedicated lane concept

Source: Cavnue

Cavnue, the state’s “master developer,” describes its approach as, “centered around creating a digital model of a roadway that analyzes and optimizes road conditions in real-time, shares information, and provides proactive guidance to vehicles and drivers. This supports enhanced safety, efficiency, and road operating environments.”

It is not clear what this means. Neither MDOT’s original request-for-proposals, nor Cavnue’s response, specify the requirements or services to be provided by a CAV corridor. Project partners are still trying to work this out. There are hundreds of open questions about what kind of data would be exchanged, over what kind of network, and for what purposes. There aren’t any existing vehicles that integrate with such dedicated infrastructure.

CAV Corridor Tollways?

There also isn’t a plan to fund the CAV infrastructure, but there are ideas for that. Public Act 179 of 2022 was enacted to support the CAV-C project and made a few changes to Michigan’s Vehicle Code. Most pertinently:

1. It authorized MDOT to designate “automated vehicles roadway lanes.”
2. It authorized MDOT to impose fees and regulate access to such lanes.
3. It authorized MDOT to contract with a third party to collect fees for automated vehicle roadway lanes.
4. It limits the liability of the automated vehicle roadway operator

In other words, PA 179 gives MDOT the ability to designate Cavnue (or anyone else) as a tolling authority. The toll facility would have to be located on MDOT trunk line but is otherwise unrestricted by this legislation. PA 179 defines “automated vehicle roadways” to include any “system that is capable of facilitating the deployment of … a vehicle equipped with varying levels of automated technology.” Automated technology can include any “technology installed on a motor vehicle that has the capability to assist … a human operator.” Since any car can be said to have technology that assists the human operator (and by a stricter definition most cars now include SAE level 1 or 2 automation), by the letter of the law, Michigan’s legislature has given MDOT authority to designate any trunk line an automated vehicle roadway and assign a tolling authority.

Michigan’s legislature has given MDOT authority to designate any trunk line an automated vehicle roadway and assign a tolling authority.

Refocus on the Real

Michigan’s CAV program, along with other current public sector efforts to prepare for autonomous vehicles, are unlikely to have any positive impact. These policies just don’t have any technologies to be applied to. It is good that Michigan policymakers are aware of the potential for this technology to be transformative, but have assumed too much. As noted by Ford CEO Jim Farley, whose company just wrote-off a $2.7 billion investment in AVs, “fully autonomous vehicles at scale are a long way off.” CAVs will likely never exist as MDOT has conceived of them. No one has figured out how to deploy AVs in meaningful numbers, and building AVs that require connectivity to expensive infrastructure is not going to help.

There should be more reflection and less reflex when investing public resources.

Michigan cannot be a leader in automotive by chasing transient technology trends. The industry appreciates that Michigan officials are interested and engaged in the development of new technologies, but there should be more reflection and less reflex when investing public resources. Such efforts distract from finding real solutions to real problems that exist today. 

Michigan’s road agencies can rest easy knowing that they will not wake up one morning to a swarm of AVs on their streets. Unfortunately, this also means that AVs aren’t going to be a solution to any of today’s problems. We have many issues to address regarding the condition, safety, and operation of our transportation system. These are hard problems. Building the infrastructure we need for the state we want to be will require focused, dedicated, long-term strategies. We cannot afford to be distracted by magic technology solutions.

March 2023 Update: Dead links updated to web archive versions.