Bavovna provides an AI‑driven navigation stack for uncrewed vehicles, fusing camera, LiDAR, and IMU data with deep‑learning perception models to deliver real‑time localization, obstacle mapping, and path planning in GPS‑limited or complex environments. The solution includes a lightweight onboard planner, a cloud‑based training pipeline that improves performance from fleet data, and a modular SDK compatible with ROS2 and PX4, enabling rapid integration across aerial, ground, and maritime platforms.
Funding
$2.7M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
1OFounders
Product
Problem
Uncrewed vehicles such as drones, autonomous ground robots, and maritime platforms often lack reliable, real‑time navigation capabilities in complex or GPS‑denied environments, leading to reduced operational efficiency and higher risk of mission failure.
Solution
Bavovna delivers an AI‑driven navigation stack that enables uncrewed vehicles to perceive, localize, and plan paths autonomously across diverse terrains. The platform fuses sensor data (camera, LiDAR, IMU) with deep‑learning perception models to generate robust obstacle maps and semantic understanding of the surroundings. A lightweight onboard planner computes safe trajectories in real time, while a cloud‑based training pipeline continuously improves model performance from fleet data. Integration is provided through a modular SDK that supports common robotics frameworks, allowing developers to embed advanced navigation without extensive custom development.
Target Audience
Primary customers are manufacturers and operators of autonomous drones, ground robots, and unmanned surface vessels seeking to enhance navigation reliability and reduce development time.
Features
- Multi‑sensor fusion engine combining visual, LiDAR, and inertial data for accurate localization in GPS‑limited settings
- Deep‑learning perception models for obstacle detection, terrain classification, and dynamic object tracking
- Real‑time onboard path planning with collision avoidance and dynamic re‑routing capabilities
- Scalable cloud training infrastructure that aggregates fleet data to refine models and reduce on‑device compute load
- SDK with ROS2 and PX4 compatibility, offering plug‑and‑play modules for aerial, ground, and marine platforms
- Edge‑optimized inference using TensorRT/ONNX for low‑latency operation on embedded hardware