Demonstrating ViSafe: Vision-enabled Safety for High-speed Detect and Avoid
Parv Kapoor, Ian Higgins, Nikhil Keetha, Jay Patrikar, Brady Moon, Zelin Ye, Yao He, Ivan Cisneros, Yaoyu Hu, Changliu Liu, Eunsuk Kang, Sebastian Scherer·May 06, 2025
Summary
ViSafe, a vision-based system, ensures safe high-speed operations for aerial vehicles. It uses a learning-based edge-AI framework with custom multi-camera hardware, proving effective in simulations and real-world tests. This system sets a new benchmark for autonomous collision avoidance in high-speed aerial navigation.
Introduction
Background
Overview of the current state of high-speed aerial vehicle operations
Challenges in ensuring safety during high-speed navigation
Objective
To introduce ViSafe, a vision-based system designed for safe high-speed operations of aerial vehicles
Highlight the system's use of a learning-based edge-AI framework and custom multi-camera hardware
Emphasize the system's effectiveness in simulations and real-world tests
Method
Data Collection
Description of the data sources used for training and testing the system
Explanation of the types of data collected (e.g., images, videos, sensor data)
Data Preprocessing
Techniques employed for cleaning, normalization, and augmentation of collected data
Discussion on how preprocessing enhances the system's performance
Model Training
Overview of the AI models used in ViSafe
Explanation of the training process, including the algorithms and datasets
System Integration
Description of the integration of the AI models with custom multi-camera hardware
Discussion on the system's architecture and components
Results
Simulation Testing
Presentation of results from simulation tests
Analysis of the system's performance under various conditions
Real-World Testing
Description of real-world test scenarios
Presentation of outcomes and challenges encountered during testing
Conclusion
ViSafe's Contribution
Summary of ViSafe's role in advancing autonomous collision avoidance in high-speed aerial navigation
Future Directions
Discussion on potential improvements and future research directions for ViSafe
Impact
Analysis of ViSafe's potential impact on the field of aerial vehicle operations
Basic info
papers
robotics
artificial intelligence
Advanced features