We've created a range of interactive agents to address different training needs, like object tracking, autonomous driving, indoor/outdoor navigation, and drone tasks. Researchers can easily choose the agent types and environments they need for training. Our platform covers six main types of agents: virtual humans, animals, cars, motorcycles, and drones.
Each agent type shares most API interfaces, including control of movement, navigation, and texture adjustments. Plus, we are developing interaction features among agents.
A diverse range of mesh bodies for virtual characters are used to construct rich and highly realistic environments. These virtual characters can be employed to create dense crowds and simulate real urban scenes, and each character can be controlled and randomly change appearance through our API interfaces.
We also have developed various types of vehicles including cars and motorcycles. These can be utilized for additional tasks such as autonomous driving and simulating dynamic traffic scenes.
The drone can be controlled like vehicles through our APIs. It can be used for multi-agent collaboration or anomaly detection tasks, etc.
