AURORA S
| ||
Overview
Introduction
AURORA S is a next-generation compact AI integrated spatial perception system that combines visual perception, inertial navigation, and self-developed AI deep learning vSLAM technology. It features high integration and plug-and-play capabilities, suitable for high-precision 3D mapping, perception, and spatial positioning in various indoor and outdoor scenarios. It can be widely applied in fields such as embodied intelligence, industrial automation, digital twins, and low-speed unmanned driving.
Core Features
- AI deep learning engine
- Provides indoor and outdoor 3D real-time map construction and positioning functions
- 6DOF spatial positioning: Provides high-precision position and attitude information in real time
- End-to-end stereoscopic depth estimation: Generate dense depth data in real-time
- AI Object recognition and segmentation: Generate item segmentation maps in real-time
- Expandable LiDAR: Provides higher accuracy 2D maps
Software and Development Support
- Paired with Aurora Remote UI customized interactive software to achieve scene reproduction and digital twins
- Provides a comprehensive Aurora Remote SDK, supporting C++, ROS1/ROS2, Python, enabling rapid secondary development, building personalized applications, and accelerating downstream product deployment
Feature Overview
- AI deep learning vSLAM engine
- Integrated design, compact size
- Ready to use, no external dependencies
- Suitable for both indoor and outdoor use, strong environmental adaptability
Aurora S System Composition
The Aurora S product consists of a binocular fish-eye camera, IMU, LiDAR (optional), and an internal computing board, providing customers with integrated 3D point cloud maps, 6DoF spatial positioning, dense depth maps, object segmentation maps, and other data. Users can reliably obtain real-time data through the Aurora Remote SDK for rapid development of applications such as dense mapping/scene reconstruction, semantic map construction, path planning and navigation, and 3DGS. Additionally, it can also quickly evaluate products and scene reconstruction through Aurora Remote UI visual interactive development software.
Origin of Body Coordinates
The positioning coordinates of the SLAM system are based on the position of the origin of the device in the world coordinate system, see the annotation of the mechanical dimension diagram for details.
Specifications
Technical Specifications
| Core parameters | Specific indicators | |
|---|---|---|
| Maximum mapping area | >1,000,000 ㎡ | |
| Repositioning | Supports global repositioning with an accuracy of ±5cm | |
| Map management | Supports map continuation, loading, and saving | |
| Mapping and positioning mode | Mainly based on deep learning's vSLAM, it can integrate LiDAR | |
| Multi-sensor synchronization mechanism | Hardware time synchronization | |
| Camera specifications | Binocular fisheye, 60mm baseline, FOV 180°, global shutter, RGB | |
| Camera frame rate | Typical 15HZ, 10/30Hz customizable | |
| Dense depth camera function | Adopts an end-to-end deep learning solution, supports high light and low texture, effective detection rate >90% | |
| AI object recognition and segmentation | Supports 18 outdoor scenes and 80 indoor scenes, expandable and customizable as needed | |
| Maximum tilt angle | Optional LiDAR | No requirements, (to ensure the 2D mapping effect, it is recommended that the tilt angle does not exceed 30 °) |
| 2D map resolution | 2cm/5cm/10cm adjustable | |
| Lidar distance measurement | Maximum range 40m@70% reflectivity | |
| Power consumption | 10W (typical, LiDAR not included) | |
| Operating temperature | -20℃~50℃ | |
| Start temperature | ≥0℃ | |
| Storage temperature | -20℃~60℃ | |
Interface Parameters
| No. | Type | Specification | Parameter/Description |
|---|---|---|---|
| 1 | Power input | XT30PW-M | DC 9~24V |
| 2 | Power input | USB Type-C | Supports USB PD3.0 protocol power input |
| 3 | Extended interface | USB 3.1 Gen1 | Supports OTG mode, compatible with Device/Host; |
| 4 | I/O interface | SH1.0-6PWB | 6Pin peripheral expansion interface for LiDAR linking |
| 5 | Data interface | Ethernet RJ-45 | Supports Gigabit Ethernet (1000BASE-T), full-duplex communication; Default IP: 192.168.11.1 |
| 6 | Movement indicator | White LED | Used for indicating the operating status of device |
| 7 | Status indicator | Red/Green two-color LED | Used for indicating device status |
| 8 | Reset button | Press and hold the power button for 30 seconds to clear the configuration of the device. After clearing the configuration, it will automatically restart | |
| 9 | Storage expansion | TF card slot | Supports TF card expansion for log storage |
Status indicator description:
| Indicator status | Description |
|---|---|
| White steady on | The system is running |
| Green flicker | Startup complete, waiting for initialization |
| Green steady on | Initialization completed, start mapping |
| Red steady on | Device exception |
Product Usage
Ethernet
The default configuration mode for Aurora S Ethernet is static IP mode with an IP address of 192.168.11.1. The computer is connected to Ethernet and 192.168.11.1 can be accessed through a browser to obtain device information for Aurora S and perform simple configuration on Aurora S.
Mapping
1. Preparation
Prepare a laptop, and install AuroraCore Remote on the computer.
2. Aurora Boot Up
Use a DC 9-24V 2A power supply to power on the device;
3. Connect Aurora
The computer connects to the Aurora device via Ethernet or by connecting to a hotspot (SLAMWARE-Aurora-xxxxxx). Hotspot connection requires the use of a driverless network card from the accessory pack to be usable; the accessory is optional.
Run Aurora Remote, enter the IP address "192.168.11.1" in the "Manually enter address" field in the pop-up window, and then click the "Connect" button to connect to the device.
After the connection is successful, the drawing screen is displayed.
4. Aurora Initialization
1. Align Aurora S towards areas with more features, within a distance of 2 meters, avoiding environments with fewer features such as open plains, refractive environments such as large areas of glass, and areas with more dynamic objects to ensure sufficient initialization features and obtain better data results.
2. Keep the device in a stationary state.
3. Click on "Device Operations" -->"Reset Map" on the Aurora Remote main interface. The device is starting to initialize.
4. After initialization is complete, an Aurora Remote pop-up prompt will appear in the lower right corner, indicating that initialization is complete.
5. Start Mapping
⚫ Route planning and suggestions
- Ensure as many observations as possible during the scanning process.
- Try to avoid scanning new areas as much as possible, and you can take a certain loop.
- Avoid the impact of dynamic objects as much as possible.
- Walk as many closed-loop loops as possible.
- Do not repeat the closed-loop area to reduce memory consumption.
⚫ Precautions for mapping
- Please clear the map before preparing to create a complete new map, otherwise the map optimization engine cannot be guaranteed to take effect.
- Keep the equipment level. Generally, the equipment should not be tilted more than 20 degrees as much as possible.
- Keep the equipment stable and avoid significant shaking. Sudden stops or movements will affect the accuracy and effectiveness of mapping to a certain extent.
- After the loop returns to the starting point, keep the robot moving, take more overlapping paths, and do not stop moving immediately.
- After returning to the origin of the loop, if the map is not closed, continue walking until the loop is closed.
- When creating maps with your hand, walk at a normal walking speed. When encountering spaces with fewer features or narrow spaces, or when turning, it is recommended to slow down.
- When scanning indoor scenes involving multiple rooms or floors, please open the indoor door in advance. When passing through the door, scan slowly and stay on the side of the door for a period of time to ensure that the features on both sides of the door can be scanned at the same time. If the door is not open during scanning, slowly turn around before approaching the door, turn the instrument away from the door, turn your back to open the door, and enter slowly.
- In and out:
- It is necessary to enter and exit sideways to ensure that the laser and vision have a common field of vision before entering, and better connect the data.
- Entering and exiting a confined space: After scanning a confined space, it is necessary to observe whether the reference objects are sufficient and whether the structural features are obvious during the scanning process.
- If the above two conditions are not met, when exiting a confined space, try to align the perspective with areas with good structured features as much as possible, while avoiding excessive perspective switching.
6. Save Map
Click on "Map Manager" -->"Download to File" on the Aurora Remote main interface to save the map as an stcm format file.
Aurora Remote Software Description
1. Main interface
Use the language switch option to select the preferred language
2. Switch perspective
AuroraRemote offers three perspectives for customers to choose from; you can switch between different perspectives by clicking the "Switch Perspective" button on the main interface.
3. Camera view
Click the "Camera View" icon on the main interface to preview the camera feed, depth camera feed, and semantic perception feed in real time.
4. IMU view
Click the "IMU View" icon on the main interface to view gyroscope and accelerometer data in real time.
5. Real scene dense mapping
Open Dense Mapping of Real Scenes-> "Select Folder...", and select the corresponding folder to store the recorded dataset.
6. Zoom/drag map
- Slide the mouse wheel to zoom the map.
- Click and hold the right mouse button, move the mouse to drag the map.
- Click and hold the left mouse button, move the mouse to rotate the map.
7. Reset map
If the mapping is completed, or you want to rebuild the map during the mapping process, you can use the "Device Operation" -> "Reset Map" function.
Resetting the map will clear the established map and reinitialize the device.
8. Synchronize map
The "Sync Map" button synchronizes the full updated map from the background to the interface. Considering that global updates are time-consuming, we only update the map incrementally near Aurora during the mapping process, and only do full updates when a closed loop is detected.
9. Save/upload map
Save as a stcm formatted file.
Click "Map Manager" on the main interface, where "Download to file" saves maps in stcm format.
Upload the map by clicking "Upload from file", and select the map file you want to upload.
The upload progress will be displayed during the upload process.
After the upload is completed, the lower right corner prompts that the upload is successful.
10. Pop-up prompt
During the operation of Aurora Remote, some messages sometimes pop up in the lower right corner of the main interface.
The main pop-up messages of Aurora Remote are shown in the table below.
| Message | Description |
|---|---|
| Initialization succeeded | Aurora S camera initialized successfully, can start mapping |
| Initialization failed | Aurora S camera initialization failed, unable to build map |
| Map optimization completed | Closed-loop conditions are detected and successfully closed-loop |
| Track lost | Feature points not detected in the camera image |
| Track lost recovery | Feature points re-detected in the camera image |
| Map loading started | - |
| Map loading completed | - |
| Map saving started | - |
| Map saved successfully | - |
Resources
SDK
Support
Technical Support
If you need technical support or have any feedback/review, please click the Submit Now button to submit a ticket, Our support team will check and reply to you within 1 to 2 working days. Please be patient as we make every effort to help you to resolve the issue.
Working Time: 9 AM - 6 PM GMT+8 (Monday to Friday)