3D Autonomous Path Planning

Take away the satellites. Land it anyway. On the spot you promised.

Evaluation only · Open

Develop a smart algorithm for autonomous navigation along a permissible corridor in 3D and implement it successfully on hardware mounted on a standard drone. The guidance algorithm must take active feedback from the ground about its tracked position and fuse it with the view from the drone camera to improve positional accuracy — and therefore the reliability of autonomous landing. The hardware must be lightweight with minimum power requirement.

Overview

Outcomes - A drone that holds a defined 3D corridor without GPS and knows it is holding it - Higher positional accuracy from fusing ground-tracked position with onboard camera vision - Autonomous landing that is repeatable, not lucky - A payload light enough and power-frugal enough to fly on a standard drone without redesigning it - A working hardware demonstration, not a simulation with confident narration Overview GPS is a beautiful thing right up until it is not — indoors, under a canopy, between buildings, inside a tunnel, near interference, or on the exact day of your demo. The problem is not just flying blind; it is flying blind along a permitted corridor in three dimensions and then putting the aircraft down accurately at the end of it. Two independent sources of truth exist: what the ground observer tracks about where the drone is, and what the drone's own camera sees. Neither alone is enough. Fusing them is the assignment. The Story The drone lifts off, confident. Twelve seconds later the GPS lock dies and it becomes, in a technical sense, a very expensive leaf. It drifts out of the permitted corridor. It descends wherever it feels like descending. Somewhere below, a person with a clipboard is writing something unkind. Fix the twelve-second mark.

The brief

Plans and holds a path along a defined permissible corridor in full 3D, not a 2D route with an altitude number stapled on
Ingests active ground feedback on the drone's tracked position as a live input, not a post-flight correction
Fuses that ground feedback with the drone camera's view to produce a positional estimate more accurate than either source alone
Uses that improved estimate to make autonomous landing reliable and repeatable at a designated point
Runs on lightweight hardware with minimum power draw, mounted on a standard drone
Behaves sensibly when an input degrades or drops out entirely, rather than confidently flying into a wall

Deliverables

The 3D corridor path-planning and guidance algorithm, with its logic and constraints documented
A sensor fusion module combining ground-tracked position with camera-derived visual positioning, and a stated accuracy improvement over each source alone
Physical hardware integrated onto a standard drone, with weight and power consumption measured and reported
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

Evaluation criteria

Positional accuracy achieved along the 3D corridor without GPS — 30%
Quality of sensor fusion between ground feedback and camera vision — 25%
Reliability and repeatability of autonomous landing — 20%
Hardware weight, power efficiency and fit on a standard drone — 15%
Live hardware demonstration quality — 10%

More challenges to enter

Prizes are offered and awarded by Software Technology Parks of India, not by Wooble. Amounts shown are as stated by the host and may include the host’s own valuation of non-cash items. Who places, how and when a prize is paid, and any tax or deduction, are between the winner and the host. A challenge may close, change or be withdrawn before results are declared.