WeRide has launched WeRide Genesis, its proprietary general-purpose simulation model that bridges physical AI and generative AI, connecting the real and simulated worlds to accelerate the development, training and validation of autonomous vehicles at scale.
Leveraging generative AI, WeRide Genesis can rapidly generate highly realistic virtual worlds, building simulated cities within minutes and accurately reproducing rare, extreme real-world driving scenarios that are difficult to capture on actual roads. This enables AVs to undergo large-scale training and validation in simulation. Insights generated are continuously fed back into real-world operations, significantly enhancing AV performance in complex scenarios, speeding up iteration cycles and reducing time and cost compared with traditional road testing.
WeRide Genesis was developed to address the challenges of commercializing autonomous driving on a global scale. Cities differ greatly in road networks, traffic behaviors, infrastructure, weather conditions and regulatory environments, placing greater demands on the generalization capabilities of autonomous driving technology. However, real-world testing is costly, slow and limited in scenario coverage, making large-scale validation across multiple cities and conditions extremely difficult. A high-fidelity simulation platform capable of fully replicating real-world environments has therefore become indispensable.
Such simulation platforms allow AVs to ‘drive’ in virtual cities, exposing AI drivers to diverse road, weather and traffic scenarios, including rare and high-risk events. This overcomes the cost and coverage constraints of real-world testing, improves training efficiency and ensures AVs are safer and more reliable on real roads.
Four core AI modules
As one of the industry’s most advanced autonomous driving simulation platforms, WeRide Genesis integrates four core AI modules – AI Scenarios, AI Agents, AI Metrics and AI Diagnosis – to handle any driving scenario, create virtual worlds in minutes and replicate real-world conditions with centimeter-level fidelity. This provides powerful support for the training, validation and iteration of autonomous driving algorithms.
• A ‘Lego’ world with unlimited combinations – Leveraging generative AI, WeRide Genesis builds highly realistic cities in minutes, accurately reproducing road infrastructure, weather conditions, environmental details and diverse traffic behaviors from around the world.
Within this ‘Lego city,’ the AI Scenarios module simulates a wide range of key scenarios that AVs could face on real roads. This includes sudden vehicle cut-ins, unprotected left turns, emergency evasive maneuvers, pedestrian or rider intrusions, and extreme events such as fires, earthquakes, road blockages and bad weather. The module draws on billions of kilometers of real-world driving data, and over eight years of long-tail edge cases from WeRide’s public road operations, ensuring AV systems are equipped to handle complex edge scenarios.
• High-fidelity interaction with agile responses – High-fidelity modeling of road users is widely regarded as one of the most challenging problems for autonomous driving simulation. The key difficulty lies in moving beyond representing an ‘average’ road user to accurately capture unpredictable behaviors in the real world – for example, a human driver abruptly cutting into the lane of an AV.
To address this, WeRide Genesis introduced the AI Agents module, which builds intelligent behavior models for human drivers, pedestrians, riders and other road users. These models realistically simulate the full spectrum of traffic behaviors, from routine driving to high-risk actions.
By combining intelligent scenario sampling with behavior modeling, WeRide Genesis can simulate a wide variety of road types, traffic conditions and the distribution of dynamic and static participants. This allows tech teams to evaluate the safety of AV decision-making and system robustness under different algorithm configurations, predict performance across multiple operational design domains (ODDs) and continuously improve iteration efficiency and quality.
• Closed-loop optimization with continuous feedback – Safety and passenger comfort remain a core priority of autonomous driving, and WeRide Genesis enhances this through the AI Metrics and AI Diagnosis modules, forming a systematic closed-loop framework for evaluation and optimization.
The AI Metrics module provides a quantitative evaluation system covering safety, compliance, comfort and travel efficiency. It translates driving behaviors into measurable performance metrics, enabling automated assessment of algorithm improvements. For example, passenger discomfort caused by sudden braking is captured as a specific comfort score and fed back to algorithm teams in real time for targeted optimization and revalidation.
The AI Diagnosis module automates root-cause analysis and provides recommendations for improvement. It can detect suboptimal driving behaviors, analyze their underlying causes and provide actionable solutions. For example, when an AV encounters perception delays or prediction errors in complex interactive scenarios, the system can quickly diagnose, fix and revalidate performance, ensuring the vehicle constantly behaves as expected.
• One platform for global use – WeRide Genesis is a versatile simulation platform that supports diverse urban road conditions, sensor perspectives and vehicle configurations. It can accommodate a range of AVs, from Level 2++ advanced driver-assistance systems to autonomous Level 4 robotaxis. This allows WeRide’s AI drivers worldwide to be trained and tested on a single platform without the need for separate simulations for different cities or vehicle types, optimizing R&D resources and accelerating AV deployment.
This technological strength has helped WeRide hold autonomous driving permits in eight countries, with AVs present in over 40 cities across 11 countries.
Acceleration flywheel
Through its four core AI modules, WeRide Genesis has built a complete closed-loop iteration system that automatically generates high-fidelity scenarios, identifies performance bottlenecks, highlights weak points and suggests improvements. This condenses millions of kilometers of road testing into a few days of virtual simulation, creating a true ‘acceleration flywheel’ for WeRide’s technology evolution, according to the company.
“WeRide Genesis builds us a digital universe that can be generated, scaled and evolved on demand,” said Dr Yan Li, co-founder and CTO of WeRide. “With it, our AI drivers can familiarize themselves with the driving environment of any city worldwide within minutes, laying a solid technical foundation for the global commercialization of autonomous driving. This represents a true leap forward in industry capability.”
In related news, WeRide launches WeChat Mini program for robotaxi service