The global technological revolution and industrial transformation are advancing at an unprecedented pace. Innovations such as big data, cloud computing, and artificial intelligence are driving autonomous driving into a new stage of development. As a core element of the digital economy, autonomous driving technology has become a strategic high ground contested by countries worldwide. Recently, China’s Ministry of Industry and Information Technology emphasized steadily advancing the industrialization of autonomous driving, injecting strong policy momentum for Chinese enterprises to seize a leading position in global competition.
In the wave of industrializing autonomous driving, Autowise.ai, one of the first enterprises to enter the driverless vehicle domain, continues to push the boundaries of technological innovation and commercial model exploration. This effort injects transformative energy into China’s progression toward a globally leading position in autonomous driving.
Overcoming Challenges in L4 Autonomous Driving
The complexities of urban traffic present significant challenges for L4 autonomous driving technology. These include the need for high precision and robustness in perception systems. L4 systems must ensure full control in specific areas and address emergencies and extreme conditions, such as obstacle detection, vehicle localization, and path planning. These demands require breakthroughs in core technologies like perception, decision-making, and control, as well as robust capabilities in technology optimization and iteration.
Technological Leadership: Building a World-Class Autonomous Driving Technology Stack
Multi-Sensor Fusion and BEV Multi-Task Perception Module
Sanitation scenarios, though seemingly simpler than urban traffic, require sophisticated L4 autonomous driving capabilities. Beyond detecting vehicles, pedestrians, and traffic signals, sanitation vehicles must address specific challenges like identifying road debris, potholes, or trash while navigating complex traffic conditions and long-tail issues. Vehicles must distinguish obstacles, assess road conditions, and execute operations with precision, such as edge-cleaning along curbs.
Leveraging expertise across diverse scenarios, Autowise.ai developed a comprehensive L4 autonomous driving stack. This system integrates high-precision maps, LiDAR, millimeter-wave radar, machine vision, and high-precision localization to enhance its perception and prediction capabilities. The in-house BEV (bird’s-eye view) multi-task perception module is industry-leading, supporting multiple tasks, modalities, and time sequences. It optimizes for sanitation-specific tasks, such as trash detection and low-lying obstacle identification, addressing unique challenges in sanitation operations. The BEV perception module uses panoramic imagery and point clouds to simultaneously predict five key tasks: 3D detection, semantic segmentation, low-obstacle detection, trash detection, and velocity prediction. This cross-temporal perception capability improves accuracy by over 16.6%, enabling precise motion prediction with an average error of just 0.29 m/s.
Emergency Handling and Precision Control
Sanitation vehicles must excel in emergency scenarios, stabilize in adverse road conditions, and execute precise operations. Autowise.ai employs 3D trajectory planning to navigate edge cleaning and incorporates vehicle-specific features into its algorithms for seamless navigation around obstacles. The system ensures precision to within 3 centimeters, delivering safety and efficiency in operations.
Remote Operations and AI Safety Assistants
Autowise.ai developed the Radar24 remote operations platform and a new-generation 1vN (one-to-many) remote operation system. By integrating multi-modal AI models, this system serves as an AI safety assistant, monitoring real-time vehicle data and alerting operators to potential risks. This advanced capability significantly enhances safety and reliability in autonomous systems.
Product Innovations: From Retrofitting to Mass Production
Autowise.ai has built a comprehensive product matrix: Autowise V3, Platforma-X, and Roboard-X, offering tailored solutions for different industry scenarios.
Autowise V3: Widely Applied in Autonomous Sanitation
Autowise V3, equipped with Sentrant’s full-stack L4 system, features cutting-edge sensor technologies like LiDAR, millimeter-wave radar, and HD cameras. Its large-capacity water and waste tanks support extended operations, delivering the workload of 2–4 manual vehicles at 40% lower costs. Deployed in over 30 global cities, it is a leading example of Chinese autonomous driving technology achieving scale in international markets.
Platforma-X: Multi-Function Autonomous Sanitation Vehicle
Platforma-X is an industry-first modular sanitation vehicle that supports sweeping, washing, and garbage transfer through interchangeable modules. Its cabless design reduces operational costs by 70%, paving the way for fully automated, cost-efficient sanitation solutions.
Roboard-X: Multi-Scenario Adaptability
Roboard-X features a shared autonomous driving chassis adaptable to passenger vehicles, logistics vehicles, and sweepers. This flexibility enhances usability and cost-efficiency while extending L4 technology to diverse applications.
