Student Projects

VE/VM450

5G Networked Drone

Instructors: Prof. Chong Han
Team Members: Yanjun Chen, Yifeng Kuai, Ye Wei, Yichi Zhang, Xiangyu Shi

Project Video

Team Members

Team Members:

Yanjun Chen, Yifeng Kuai, Ye Wei, Yichi Zhang, Xiangyu Shi

Instructors:

Prof. Chong Han

Project Description

Problem

Current drone communication system faces two main limitations. The drones have to transmit information by the point-to-point Radio Frequency (RF) signal. A remote controller is always required even if a mobile App is applied. Most drones can only be controlled within 5 kilometers. Another issue is that it will be difficult for supervision, if all drones are connected locally. This project aims to use DJI OSDK and network programming to solve the distance and supervision problem.

Fig. 1 Current drone communication system [1]

Concept Generation

Software Development Kits (SDK) should be used to directly control the motion of the drone. A control terminal should be developed to replace the traditional remote control. A communication system supporting long-distance control (>5km) is required as well.

Fig. 2 Concept diagram of our project [1][2]

Design Description

Our design uses an Android App, an FRP server, and an onboard computer, and combines them together. The onboard computer is automatically connected to the FRP server, and continuously receive commands from it. When the user presses a button on the App, control signals are sent to the server, then to the onboard computer. After that, the onboard computer transforms the signal to flight control signals including takeoff, landing, hovering, etc. Similarly, the onboard computer sends back drone information (GPS data, battery power, etc.) when flying.

Fig.3 Overview of our network system

Modeling and Analysis

We use DJI simulator to simulate how our drone actually works. When using the DJI simulator, the drone needs to be connected to our computer, so the result is very reliable. By DJI simulator, we can check many aspects of our design, such as takeoff and landing of our drone.

As long as the drone has no hardware problems, the performance of the drone in reality should be the same as the simulation results.

Fig.4 Drone taking off in the simulator

Validation

Validation Process:
The validation is divided into 3 stages:
Simulation Test
DJI simulator is used to verify that the system communication is successful. Latency and transmission speed are checked at this stage.
2. Prototype Test
Huawei 4G module is connected to a real DJI M600 Pro drone. We control the drone from our App through the network we built, to verify the real-time control.
3. Prototype Remote Test
To test the control distance, we do the same test in stage 2 with the App terminal on Xuhui Campus, and the drone on Minhang Campus.
Validation Results:
According to validation part, most specifications can be met.
√ Distance from the user>=10km
√ Data transmission latency<=100ms
√ Max transmission bitrate>=1Mbps
√ Visualization units in the UI>=3

Conclusion

We constructed a communication system between the DJI Manifold, the drone and the Android phone. Also, we used a 4G module to connect the DJI Manifold to the internet. Through our network, we can control the drone by our Android App.

Acknowledgement

Sponsor & Mentor: Jianrong Xiao from UM-SJTU Joint Institute

Reference

[1] https://store.dji.com/cn/product
[2] https://www.huawei.com/cn

UM-SJTU JOINT INSTITUTE