SKY VYOM
Developing an affordable GPS-free UAV system that uses LiDAR-based 3D mapping, obstacle detection and intelligent path planning to navigate challenging environm
- Problem, approach and result explained
- Expert reviewed· 6 reviews
—
97
—
99
—
87
Overview
The problem
GPS-based navigation is widely used in drones, but GPS signals can become weak, unavailable or unreliable in indoor environments, tunnels, narrow corridors and certain disaster-affected areas. In these situations, drones may struggle to understand their surroundings, identify obstacles and follow a safe navigation path.
For example, during flood rescue, building inspection or search-and-rescue operations, a drone may need to navigate through confined spaces where GPS-based positioning is not dependable. Without accurate environmental information and reliable obstacle detection, safe navigation becomes a major challenge.
Our project focuses on addressing this problem through a 3D Autonomous Path Planning System. We aim to develop a drone-based solution that can scan its surroundings using LiDAR and distance sensors, create a 3D environmental representation, identify obstacles and generate suitable navigation paths without depending on GPS.
We also identified another challenge: collecting sensor data is not enough. The system needs to process, visualize and communicate that information in real time so that the operator can understand the environment and monitor the planned route.
Our central problem statement is: How can a drone understand its surrounding environment, detect obstacles and plan a suitable navigation path in GPS-denied environments using affordable sensors and embedded hardware?
The approach
Our approach is to develop an affordable GPS-free 3D mapping and path-planning system using embedded hardware, LiDAR sensing and wireless communication.
We use two ESP32 development boards. The first ESP32 acts as the main sensing controller, connected to a Benewake TF-Luna LiDAR sensor and an MG90S servo motor. The servo rotates the LiDAR sensor to scan the surrounding environment at different angles, while the LiDAR measures obstacle distances. Additional GY-53/VL53L0X and HMC5883L sensors provide supplementary distance and heading information.
The second ESP32 handles wireless communication and data transfer between the sensing hardware and our web-based UAV Mission Control System. The website receives sensor readings, displays scanning information, visualizes the mapped environment and allows users to create mission waypoints.
Our software approach includes processing sensor measurements, converting scan data into local coordinates, displaying mapping points in a 3D environment and generating waypoint-based routes. The mission planner is designed to identify potential obstacles and support route planning through a local XYZ coordinate system without depending on GPS coordinates.
We are developing the system step by step, beginning with individual sensor testing, servo scanning, ESP32-to-ESP32 communication, live data visualization and mission waypoint transfer.
The RadioLink CrossFlight V1.0 remains responsible for flight stabilization. At the current prototype stage, our focus is on mapping, mission planning and hardware communication rather than claiming fully autonomous flight.
Our key innovation is combining low-cost embedded hardware, LiDAR-based environmental scanning, two-way wireless communication and a custom 3D mission-planning interface into one integrated UAV prototype.
The result
Our project has progressed from an initial concept to a working hardware prototype with individual sensor testing and system integration underway.
We have successfully tested the MG90S servo motor and Benewake TF-Luna LiDAR sensor with the ESP32. The servo can rotate through the programmed scanning angles, while the LiDAR provides distance measurements and supports maximum-distance recording.
We are developing the two-ESP32 wireless communication system to transfer scanning data between the main sensor controller and the visualization controller. Our next integration milestone is displaying live sensor readings and scanning information on our custom UAV Mission Control Website.
The mission-planning interface is being developed to generate waypoint-based routes, store mission information and visualize planned paths using local XYZ coordinates.
The project demonstrates the practical integration of embedded electronics, distance sensing, servo-based scanning, wireless communication and 3D visualization. Further testing is planned to evaluate mapping accuracy, obstacle detection, communication reliability, positioning performance and autonomous navigation.
Our final goal is to demonstrate a complete GPS-denied corridor navigation system with measurable performance through recorded hardware and flight trials.
Reflection
Reflection
Throughout this project, we have learned that developing a GPS-free autonomous navigation system is much more challenging than simply connecting sensors and writing code. The biggest lesson has been understanding how individual hardware components must work together as one reliable system.
During development, we faced challenges with sensor communication, servo movement, power supply requirements, data processing and wireless connectivity. By testing each component separately and gradually integrating them, we improved our understanding of embedded systems, LiDAR technology, wireless communication and 3D mapping.
One important learning was that distance measurements alone are not enough to create a reliable 3D map. The system also needs accurate scanning angles, coordinate calculations and reliable position information. We recognized that our current single-servo scanning arrangement has limitations and that further development is required for complete 3D environmental mapping and autonomous navigation.
We also learned the importance of building within limited resources. Instead of depending on expensive commercial systems, we are exploring an affordable solution using ESP32 boards, LiDAR sensors and a custom mission-control website.
Our next steps are to improve mapping accuracy, develop more reliable sensor fusion, integrate the hardware communication system, test obstacle-aware path planning and conduct controlled navigation experiments.
This project has taught us to approach engineering challenges through experimentation, testing, teamwork and continuous improvement. We see this prototype not as a finished autonomous drone, but as an important step towards developing an affordable GPS-denied UAV navigation system.
Submitted to 3D Autonomous Path Planning
- Your idea in one line
A Gps free 3d navigation system that uses ai , cameras , and sensors to map environments ,avoid obstacles and plan safe paths for drones .
- What will be working by the finale?
A working prototype demonstrating gps free localisation 3d mapping , autonomous path planing and obstacles avoidance in a simulated environment
- Tech stack and hardware you plan to use
Python, c++ , Ros 3/ Esp 32, camera , drone and laptop and some electronic components
- What have you built or tested so far?
We have developed a initial concept and identifies the required technology
- Pitch deck (PDF)
- Pitch deck (PDF)PDF
- The 3D corridor path-planning and guidance algorithm, with its logic and constraints documented
The 3D Corridor Path-Planning and Guidance Algorithm is designed to help the UAV understand its surrounding environment, detect obstacles and identify a safe navigation path. The system uses a TF-Luna LiDAR sensor mounted on a servo motor to scan nearby objects and collect distance and angle measurements. The ESP32 processes the collected data to generate environmental mapping points and construct a virtual 3D corridor between the starting position and the target destination. The algorithm continuously checks for obstacles within the planned corridor and calculates an alternative safe path whenever an obstruction is detected. The processed mapping data, obstacle information and planned route are transmitted through Wi-Fi to the website for real-time 3D visualization and mission guidance. The algorithm considers sensor range, scanning angle, corridor width, obstacle clearance, data update rate and position-estimation accuracy as its main constraints.
- A sensor fusion module combining ground-tracked position with camera-derived visual positioning, and a stated accuracy improvement over each source alone
The sensor fusion module combines ground-tracked position data from the UAV system with camera-derived visual positioning from the ESP32-CAM to estimate the drone's relative position and movement. The ground-tracking data and visual observations are processed together to reduce individual measurement errors and improve position estimation compared with using either source alone. The module is designed to provide more stable and reliable positioning for 3D mapping and GPS-free path planning. Accuracy improvement will be measured through testing against a known reference position; no numerical improvement should be claimed until experimental results are available. Important: Our current hardware does not yet include a complete ground-tracking system or a validated visual-positioning algorithm, so this is a planned development feature, not a currently demonstrated capability.
- Physical hardware integrated onto a standard drone, with weight and power consumption measured and reported
The UAV 3D Mapping and GPS-Free Path Planning System is being integrated onto a standard drone platform, incorporating two ESP32 boards, a Benewake TF-Luna LiDAR sensor, MG90S servo motor, GY-53/VL53L0X distance sensor and HMC5883L magnetometer. The hardware is designed to operate as an onboard sensing, scanning and communication system, providing real-time sensor data to the ground-based Mission Control Website through Wi-Fi. The complete integrated prototype will be weighed, and its power consumption will be measured under operating conditions to evaluate payload capacity, battery requirements and flight suitability. Actual weight and power consumption figures will be reported after physical testing.
- A live or recorded flight demonstration of GPS-denied corridor navigation ending in an autonomous landing, with landing error measured across repeated attempts
A live or recorded flight demonstration of GPS-denied corridor navigation ending in an autonomous landing, with landing error measured across repeated attempts
- A failure and degradation analysis — what happens when ground feedback drops, when the camera view is obscured, when both degrade at once
The project aims to demonstrate GPS-denied corridor navigation using the UAV's LiDAR-based obstacle detection, sensor fusion and 3D path-planning system. During the demonstration, the drone will navigate through a predefined corridor without relying on GPS and attempt an autonomous landing at a designated landing point. The demonstration will be recorded to evaluate navigation performance, obstacle avoidance and landing accuracy. Multiple flight trials will be conducted, and the landing error will be measured in centimetres to assess repeatability and system reliability. Actual flight footage and measured results will be documented after successful testing. Note: This is a planned validation milestone, not a completed capability of the current prototype.