Hi, I'm a final year mechanical engineering student from a University. My project was about integrating LiDAR with the tello so it can navigate autonomously whether indoor or outdoor. My question is, is there already a project where people combining the LiDAR sensor with the drone and integrating AI and machine learning so it can navigate autonomously and avoid obstacles? If so may I contact any of you for further questions?
Tello RMTT with LiDAR and Autonomous flying with Machine Learning
- Thread starter takeo
- Start date
You should take into account that the Tello has a very limited capability to carry extra payload. A Raspberry Pi Zero W + lightweight single beam LiDAR like the VL53L0X (or VL53L1X) or may be a Benewake TF-Mini may already bring the Tello to it's limit.
Another option may be the Tello Talent (or Tello EDU with the Tello Talent Expansion Kit). This contains an ESP32 and as far as I remember, there is also a LiDAR add-on available (probably also based on the VL53L0X).
The problem is, that a single beam LiDAR will not bring much information about your environment, but may be, if you rotate it continously around it's yaw axis...
Another option (but without LiDAR) would be, to use the camera stream and the API of the Tello (using a library like djitellopy) from a ground station, controlling the Tello movements. You can apply visual ORB SLAM to the video stream or a mono-depth AI model (or a combination of both).
Another option may be the Tello Talent (or Tello EDU with the Tello Talent Expansion Kit). This contains an ESP32 and as far as I remember, there is also a LiDAR add-on available (probably also based on the VL53L0X).
The problem is, that a single beam LiDAR will not bring much information about your environment, but may be, if you rotate it continously around it's yaw axis...
Another option (but without LiDAR) would be, to use the camera stream and the API of the Tello (using a library like djitellopy) from a ground station, controlling the Tello movements. You can apply visual ORB SLAM to the video stream or a mono-depth AI model (or a combination of both).
Last edited:
It should be possible to continously rotate Tello to get a 360 scan from a single beam sensor. The camera images will be mostly useless depending on rotation speed
The heading reported remains amazingly accuracte and consistent over time. You just have to constantly adjust pitch & roll for current heading and Tello will fly a straight line while spinning. (the "headless" control option in TelloFpv basically does that)
The heading reported remains amazingly accuracte and consistent over time. You just have to constantly adjust pitch & roll for current heading and Tello will fly a straight line while spinning. (the "headless" control option in TelloFpv basically does that)
I'm very interested in the ORB SLAM method. Could you recommend a guide on how to connect Tello's API to the ORB SLAM and an AI model? Additionally, where can I find the necessary code to learn with proper guidance?
I'm back again, this time i encountered a problem where I seem can't to connect my Tello to my VirtualMachine (Oracle VBox) running Ubuntu 20.04 desktop from my Windows laptop
I've been following this steps to run ORB SLAM and control it via a joystick/controller: Operate DJI toy drone “Tello” with ROS and run ORB-SLAM3
The website is in Japanese and I translated it via google translate.
Is it actually possible to connect my Tello in the VMs? I can't seem to find any solution so far.
I've been following this steps to run ORB SLAM and control it via a joystick/controller: Operate DJI toy drone “Tello” with ROS and run ORB-SLAM3
The website is in Japanese and I translated it via google translate.
Is it actually possible to connect my Tello in the VMs? I can't seem to find any solution so far.
