Automated Parking: Valet Done from Your Phone

Author: Maite Bezerra

We all dread having to find an empty parking space and perfecting vehicle maneuvers to avoid accidents. It brings great anxiety and cuts into the limited amount of time we have in a day. Fortunately, automatic parking can relieve a driver of these concerns. Cars with park assists have come a long way, from basic parking assist features to fully automated valet parking (AVP). Indeed, a self-parking car makes peoples’ lives much easier and is more convenient. These solutions even open the market for cars that can parallel park themselves.

Chart Source: ABI Research

 

As a full L4 capability, this form of automatic parking lets drivers drop off their cars and use a mobile app to let the autonomous car find an empty parking spot. And when the user is ready to go home, the car is then summoned to a pickup spot via the mobile app. Because of the low speed and low safety risk associated with parking, it’s considered the most straightforward L4 use case. The three types of automated valet parking are Type 1 (vehicle-based), Type 2 (infrastructure-based), and Type 3 (hybrid).

Get the Research Report: Parking Assistance and Autonomous Valet Parking Systems

Why Is Automated Valet Parking Worthwhile to Businesses and Consumers?

Automated valet parking has the potential to be a big winner in the Business-to-Business (B2B) space, as car rental companies can reduce the turnaround time for vehicle returns, servicing, and parking for future rentals. Plus, car park/parking garage operators will give a warm welcome to the 20% more vehicles that are promised with fully automated parking. On top of that, car factories and logistics centers can use these technologies to move cars from one production point to the next. This eliminates the need for human drivers, while also minimizing the risk of damage.

In the Business-to-Consumer (B2C) space, AVP excels at delivering convenience to users. It saves them the time and associated risk of finding an empty parking space. According to international insurer Allianz, about 40% of vehicular accidents happen when drivers are trying to park or maneuver, but automated parking can significantly cut back on these occurrences. All in all, the B2C vertical will be a very small market opportunity for the foreseeable future due to associated costs, low consumer perceived trust in automation, and carmakers focusing on B2B use cases.

The Intelligent Vehicle Approach to Automated Valet Parking

Type 1 AVP is where all the intelligence comes from within the vehicle itself, including sensors, wide-angle cameras, radars, front cameras, computing units, algorithms, etc.

How Does It Work?

With this approach, the vehicle searches for and selects an empty parking space autonomously using onboard sensors and a car park map. Then, once the vehicle approaches the space, the parking maneuvers are 100% automated and the driver gets a notification once the process is completed. Although 100% of the intelligence stems from the vehicle, Type 1 AVP still requires some infrastructure to be in place for it to work (i.e., drop-off and pick-up zones, artificial landmarks, and High-Definition (HD) maps). Parkopedia is an important enabler of autonomous parking as the company creates HD maps of off-street car parks that support indoor navigation and localization.

What Is the Market Outlook?

Type 1 is considered to be the most scalable solution to autonomous valet parking because of the minimal infrastructure requirements. Based on qualitative interview feedback, it’s favored by the car park/parking garage industry, consumers, automakers, and Tier One/Tier Two suppliers. But the most important thing right now is for proper automated parking garage infrastructure to be put in place if consumers and car Original Equipment Manufacturers (OEMs) are going to adopt the technology.

Tackling Automated Valet Parking with Infrastructure

For Type 2 automatic valet parking, all the intelligence comes from the cark park/parking garage infrastructure. Some of these infrastructure technologies include Light Detection and Ranging (LiDAR) or stereo cameras, a cloud back end, a parking garage server, and connectivity hardware.

How Does It Work?

After a driver exits the vehicle and activates the service via a smartphone app, a computer in the parking facility obtains data from several sources to determine what steering throttle and brake commands to send to each vehicle. If, for example, a pedestrian walks out in front of a vehicle, a dedicated control center will ensure that the car comes to a complete stop.

What Is the Market Outlook?

Due to significantly high installation and maintenance costs, Type 2 AVP will not be adopted in the consumer market to a meaningful degree. There also isn’t much incentive for park operators given that they likely won’t see a Return on Investment (ROI). That’s because they almost never own the buildings where they operate and ROI won’t be generated before a 5-year contract is up.

Finding a Middle Ground

Popular in the Japanese automotive industry, the Type 3 approach to AVP is a split between Type 1 and Type 2, meaning it relies on vehicle intelligence and infrastructure intelligence alike.

How Does It Work?

In this approach, parking navigation is supported by onboard vehicle sensors that work hand-in-hand with infrastructure intelligence. Localization, which is critical for L4 autonomous parking, is complicated to implement effectively. NXP has shown that Ultra-Wideband (UWB) is an effective solution as it provides spatial awareness to cars and smart devices with highly precise localization capabilities. To illustrate, UWB anchor points would be installed every 10 to 20 meters inside a facility. These anchor points would communicate with the UWB-enabled vehicle to calculate the distances of the moving vehicle via time of flight—thereby enabling autonomous vehicles to move to free spots and pick-up areas.

What Is the Market Outlook?

While the hybrid approach to AVP requires fewer onboard sensors and processing capacity for carmakers to worry about, there’s no telling how widespread UWB technology will be in the car park industry years in the future. And due to high investment costs associated with UWB, car park/parking garage operators will be hesitant and large-scale deployment is not likely. All in all, the hybrid approach is viewed as the most complex of the three approaches and demands strong collaboration among automotive ecosystem players.

The Consumer Market Is a Tough Nut to Crack

Each approach to AVP comes with its own set of pros and cons, but the vehicle-based approach (Type 1) clearly has the least roadblocks in the path to reaching the mainstream. As long as the industry swiftly identifies minimum infrastructure and/or vehicle-based localization requirements, the solution should have a bright future. Because if the automotive industry can’t agree on minimum requirements and the proper infrastructure isn’t enabled, advanced features like additional learning, sensors, and processing will be required inside the vehicle—keeping the Manufacturer's Suggested Retail Prices (MSRPs) firmly in the premium segment.

On account of the interviews ABI Research analysts have had with industry players, the infrastructure approach (Type 2) to AVP is not likely to gain traction in the consumer markets, as car park/parking garage operators aren’t inclined to invest in automated infrastructure while already dealing with tight margins. On the other hand, car manufacturing plants will find refuge in the infrastructure-based approach to automated parking to save on costs associated with human workers. Companies like Continental, BMW, Mercedes-Benz, Ford, and VW have all shown interest in this vertical and some even gave demonstrations at IAA Mobility 2021.

Finally, the hybrid approach (Type 3) is seen as the most convoluted approach to self-parking vehicles. Not only do carmakers and infrastructure players have to come together to design platforms and enabling technologies, but UWB has too many question marks surrounding its adoption. Moreover, there is also no definitive guidance as it relates to what intelligence should be in the vehicle and the infrastructure. Rather than using the reliable, yet pricey, UWB technology, market players hint at a preference for Vehicle-to-Everything (V2X) for communication between vehicles and infrastructure.