Home-zone Parking Assist (HPA) and Automated Valet Parking (AVP)

Home-zone Parking Assist (HPA) and Automated Valet Parking (AVP) represent higher levels of automation within advanced driver assistance systems (ADAS). Together, they mark the shift from drivers searching for parking spaces to vehicles finding their own parking bays. However, fundamental differences remain in their underlying technical principles and operating environments.

Home-zone Parking Assist (HPA)

HPA is a driver assistance feature built on route learning and replication. The system learns and maps out the driver's routine parking trajectory via onboard sensors. Once the initial "training" route is recorded, the vehicle can automatically replicate the path under identical conditions, taking over low-speed cruising and parking manoeuvres over the final stretch in the car park.

Core Technical Principles

HPA is powered by Simultaneous Localisation and Mapping (SLAM) technology. During the maiden learning drive, the system utilises onboard sensors to map the car park layout and establish real-time vehicle positioning. On subsequent visits, the system simply retraces the memorised route to park the car. Because HPA functions purely as a route replication feature, it does not rely on high-definition (HD) maps. It is best suited for frequent, highly repetitive scenarios such as assigned residential parking bays and dedicated office car parks.

Key Features and Industry Landscape

HPA is rapidly moving from technical validation to mainstream production. Through an over-the-air (OTA) update, the Lynk & Co Z10 introduced an HPA system capable of memorising routes up to 2 km with continuous multi-storey navigation, manoeuvring effortlessly through tight ramps and narrow lanes with just 80 cm of clearance beyond the vehicle width. Jingwei HiRain's latest mass-production HPA setup goes a step further, supporting single memorised routes of up to 3,000 metres and storing up to 10 distinct park-in/park-out paths. By leveraging visual landmark positioning and inertial navigation, it can execute pinpoint parking even in underground car parks devoid of GPS reception.

Important Note: HPA remains, at its core, a driver assistance system. Buyers often confuse HPA with fully fledged AVP, assuming HPA can autonomously search for vacant bays without human intervention. Automakers clearly specify that while HPA memorises and retraces pre-recorded paths, it does not autonomously scout for lots, and the driver is still required to remain in the driver's seat throughout the manoeuvre.

Automated Valet Parking (AVP)

AVP is regarded as the pinnacle of smart parking technology, operating as a Level 4 low-speed autonomous driving system. The driver can simply alight at a designated drop-off point and initiate parking via a smartphone app. The vehicle will then independently cruise through the car park to locate and enter an available bay without requiring any driver intervention or oversight. For the return journey, the vehicle can summon itself back to the pick-up point on demand.

Technical Approaches and Real-World Deployment

At present, AVP development is split across three main architectural routes:

• Vehicle-centric approach: The vehicle relies entirely on its suite of onboard sensors and computing algorithms to complete the entire parking cycle, requiring zero infrastructure upgrades in the car park

• Infrastructure-led approach: Dedicated sensors and cameras installed throughout the car park handle environmental sensing and bay allocation, while the vehicle merely executes driving commands

• Vehicle-to-Infrastructure (V2I) collaborative approach: The vehicle and car park exchange data wirelessly to secure real-time bay availability and optimal routing in advance, enabling seamless and highly efficient parking

The P7 Car Park at Nanjing South Railway Station stands as a benchmark for commercial AVP deployment in China. The facility underwent a comprehensive smart overhaul with an integrated digital management platform, interfacing directly with vehicle smart-driving apps. Compatible vehicles can perform fully autonomous parking and retrieval across any section of the car park, with initial support rolled out for more than 20 models, including the AITO range. In Chengdu, e-Park Technology has partnered with Dongfeng Motor to integrate in-house AVP solutions into mass-production vehicles, accelerating the commercial rollout of driverless valet parking.

Comparison Summary