Hi! I’m going to design an autonomous flight drone which aim is to navigate both indoor/outdoor without GPS. I’m looking for different possible setup of sensors in order to navigate, map and avoid obstacles. Do you have any suggestions? (My budget is not too high - I want something that works well but not the best)
@Mratoz If budget is low and time is infinite, a vision system is always about as good as it gets, but is quite complex and there’s almost nothing “out of the box”. You might be able to run OpenCV on an SBC on the drone. Outdoors - GPS has proven quite reliable, but not for tight spaces. Indoors, you might consider SLAM (again, lots of programming), or 3D Lidar / 360 degree rangefinder. There is not much when it comes to “ready out of the box” indoor navigation, even for 2D ground-based robots, let alone 3D. I have not seen a drone use a system like this, and not sure what interference the motors and noise might cause, and it’s also not “inexpensive” but good to know about…
@cbenson First of all, thanks for your very fast answer.
Yes, I’d like to consider SLAM. Thus I need some sensors for obstacle avoidance like ultrasonic and laser, and also a camera in order to understand where I need to land for example (I’m talking about a quadcopter). The problem is that I don’t know which is the best combination (laser for height detection, ultrasonic under each motor in order to cover at 360° the obstacles, etc…).
P.s. I don’t want to put sensors in the room that I’m mapping to find the position of the drone.
SLAM normally requires more than 1D sensors. Looking up the keywords Drone Indoor SLAM in YouTube, you’ll get an idea of how others have tackled the challenge. For example,
If you’re considering landing, then there are work-arounds like adding a distance sensor to each “foot” to ensure a flat surface. Might want to make this post public as others might be able to provide firsthand build experience.
Hi @cbenson thanks very much for the quick answer. Yes make this post public.
Good idea to put distance sensor to each foot.