Robotaxi: PonyPilot+
Robotaxi is an autonomous driving travel service that has transitioned into a fare-charging commercial model. The development of robotaxi is closely intertwined with the ongoing iteration of advanced autonomous driving technology, and has entered the stage of testing and operation of fully driverless vehicles. Today, residents of Beijing, Shanghai, Guangzhou, and Shenzhen enjoy using PonyPilot+ and are experiencing the joy of autonomous driving. Let’s hear their stories.
Getting Around Effortlessly
Place an order, start your trip, reach your destination – autonomous driving fuels our imagination of future travel that is more friendly and secure. Turn dull car cabins into your “mobile life space” where you can listen to music, admire the scenery, video chat with family and savor your time on the road.
Proven Strengths Bring Peace of Mind
A dependable service is a testament to the technical team’s solid accomplishments. We have honed skillful driving techniques from complex scenarios, devised fully redundant software and hardware systems, developed an autonomous remote assistance platform, introduced an autonomous driving fleet management platform, and implemented automated service experience management. Our unwavering focus on safety, comfort, and efficiency is evident in every detail of our travel services.
Next Generation Robotaxi Vehicles
Introduced in January 2022, our sixth-generation L4 autonomous driving software and hardware system was initially installed in Toyota’s Sienna Autono-MaaS vehicles. Starting in 2023, the system will be rolled out in major cities across the country, allowing us to build a fully driverless robotaxi service at scale. With a design that is nearly on par with consumer-grade vehicles, our new robotaxi vehicles have been engineered for auto-grade mass production in all aspects of hardware R&D, component selection, safety redundancy, system assembly and production.
-
x4
Roof-mounted solid-state LiDAR sensors
-
x2
Blind-spot detection LiDAR sensors
-
x3
Long-range millimeter-wave radar sensors
-
x11
Cameras
Robotruck: PonyTron
The fusion of autonomous driving and traditional logistics is opening up a realm of new possibilities for the industry. The logistics industry has been grappling with persistent issues such as labor shortage, decentralized management, and rudimentary business practices. Autonomous driving and intelligent logistics platform technologies are pivotal in empowering the logistics ecosystem of the future.
Safer
- The use of autonomous driving on highways guarantees a high level of security thanks to multi-redundant safety mechanisms that operate in all complicated weather and driving conditions.
- Compared to human drivers, self-driving cars have an accident rate per million kilometers that is ten times lower.
More efficient
- Longer mileage: All-weather operation around the clock
- Faster speed: Can maintain the max flow of traffic speed while driving
- More streamlined scheduling: robotrucks can serve as mobile terminals, enabling centralized and intensive coordination at a larger scale
More affordable
- Driver labor costs reduced by at least 40%
- Improved fuel efficiency, with fuel savings 10% greater than that of experienced drivers
- Lower accident rates lead to a drop in insurance premiums of at least 10%
“Golden Triangle”Creating a New Pattern of Logistics Together
Based on industry demands, we collaborate extensively with autonomous driving technology providers, truck manufacturers, and logistics service providers to establish the first "golden triangle" in the trucking industry. This collaboration aims to create an intelligent closed-loop logistics ecosystem that realizes the vision of a “smooth flow of transportation”.
PonyTron
SANY Group
Sinotrans Limited
RobotruckGolden Triangle
Self-Driving Truck System Solutions
Our latest self-driving truck system, specifically developed for truck logistics, was launched in November 2022. This third-generation system offers L4 autonomous driving capabilities on busy highways and expressways. The first commercial use of this system was in the first-ever self-driving heavy-duty truck co-created by Pony.ai and SANY Heavy Truck. Yiji Smart Truck (the JV between Pony.ai and SANY Heavy Truck) is responsible for the production of trucks, and finished products have been delivered to Cyantron (the JV between Pony.ai and Sinotrans). With an ultra long-range sensing module and a 360° panoramic sensing module, the sensor system is capable of detecting obstacles, such as stalled vehicles, well in advance. This ensures the safe navigation of obstacles within 200 meters around the car. The complete sensor system comes with an exclusive Sensor Arc design that integrates all sensors into a single, curved module. Additionally, a sensor cleaning system has been included to boost sensitivity and reduce maintenance complexity.
Panoramic sensing module
-
x1
360° LiDAR
-
x3
Blind spot detection LiDAR
-
x3
Camera
-
x1
Millimeter wave radar
Ultra long-range sensing module
-
x1
Super long distance camera
-
x2
Long distance camera
-
x1
Auto-grade long-range LiDAR
-
x1
Long-range millimeter wave radar
Historic Moments
Our robotruck technology made history in China in December 2021, when we conducted the first test of an autonomous truck on a public highway. This accomplishment represents a major policy breakthrough and a critical step forward for the future of intelligent logistics services. Since 2021, we have been granted the first road test permits for self-driving trucks and road transport business licenses for freight in different regions, including Guangdong and Beijing.
Personally Owned Vehicles (POV)
A confluence of factors, notably well-developed advanced hardware and a shifting consumer landscape, has paved the way for a transformative phase in our sustained investment for mass production. In January 2023, we officially announced three intelligent driving product lines, encompassing software system solutions, domain controllers, and closed-loop data toolchains. These products have received orders and are now in mass production. Drawing on our six years of developing and refining “virtual driver” technology, we are pleased to offer our customers high-performance, cost-effective bespoke intelligent driving solutions that are suitable for mass production.
Shitu
Smart Driving Solutions
We are thrilled to announce the release of three smart driving solutions – PonyClassic, PonyPro and PonyUltra – that deliver exceptional driving and riding experiences. These solutions come with various hardware configurations that set them apart, and mark the culmination of our many industry-leading technological advancements. The official launch is scheduled for 2023.
BEV Sensing Algorithm
- Large model to identify various types of obstacles, lane lines, passable areas, etc.
- Navigation maps are used to enable highway and city NOA.
- Minimized computing power requirements.
Game-Based Interactive Planning Algorithm
- Agile travel through urban elevated roads.
- Takeover metrics are multiple times better than mainstream products in the industry.
A Seamless Driving and Riding Experience
- Pioneering all-scene ACC, all-scene LCC, all-scene NOA.
- Consistent algorithms for the same function in different driving scenarios to ensure a coherent experience.
Fangzai
Autonomous Driving Controller (ADC)
With a focus on performance, cost and reliability, our auto-grade domain controller Land (Fangzai) is a true fit for intelligent driving. By fully understanding the needs of autonomous driving hardware design and software adaptation, we have made Land (Fangzai) more compatible with upper-level software. We have achieved improved operating efficiency and versatility for use in various scenarios, such as urban driving, unmanned delivery, environmental sanitation, ports and mining areas.Land (Fangzai) currently uses an autonomous domain controller powered by NVIDIA DRIVE Orin, which provides optimized hardware performance. We are actively working on adapting to other chips and domain control platforms, as well. The domain controller has been certified for compliance with the automotive functional safety ASIL-D development process and international quality system standards.
250 × 182 × 31 mm
Single NVIDIA DRIVE Orin (Liquid cooling)
250 × 193 × 64 mm
Single NVIDIA DRIVE Orin (air cooling)
306 × 205 × 45 mm
Double NVIDIA DRIVE Orin (Liquid cooling)
Autonomous Driving Functions
- Active safety
- Driving: Highway NOA, City NOA
- Parking: Memory parking, Valet parking
Use cases
- Intelligent driving assist
- Low-speed autonomous driving
Cangqiong
Data Closed-loop Toolchain
The closed-loop data toolchain plays a vital role in the development and success of intelligent driving systems. The toolchain product Sky (Cangqiong) was launched to serve auto manufacturers and Tier 1 customers, assisting them in leveraging the data and rapidly improving the capability of intelligent driving systems. Sky (Cangqiong) covers two core requirements: one is the closed-loop full volume data in the R&D and testing phases, and the other is the shadow mode-based closed-loop data in the mass production phase.
Addressing the issue of the rapid growth of data
Precision evaluation system capability
Vehicle-cloud coordinatedbig data platform
- Data collection
- Dataset management
- Data mining
- Data analytics
Cloud-based large-scalesimulation platform
- Participants of intelligent transportation
- Perfect consistency of behavior
- Smart task scheduling