Drone patrol: detailed report
In our last publication, we touched upon the topic of monitoring with the help of drones and showed the first tests of the Drone Patrol project. In this article we will describe in detail the results of our experiments on the applicability of drones for monitoring open areas with a large territory and hard to access topography.
In all experiments, the following equipment and software were used:
- Quadrocopter DJI Matrice 100
- Stabilized camera DJI Zenmuse Z3, 12.4 Mpx
- Drone Employee DJI SDK — our software for controlling the DJI drones
We did not consider the issue of the initial receipt of the alarm signal for the launch of the drone. It is assumed that information about a possible incident is already known and transferred to the base.
Experiment over the landfill for solid domestic waste
Venue of the experiment: “Zhigulevskaya Valley” Technopark, Togliatti, Russia — conserved landfill.
Togliatti is a large industrial and economic center. There is an increased number of landfills for storage of waste within the city limits. At the last stage of their life they are eliminated by the method of conservation — the mass of garbage is buried underground and covered with protective material. This allows further reclamation of the territory of the landfill for reuse. However, the conserved landfills polygons attract the attention of so-called “black diggers” — people involved in the illegal digging of dumps for the search for non-ferrous metals. They re-open the landfills, which ultimately leads to negative consequences for the environment and the population. Revealed pollution threatens the animal and plant life, waste enters the groundwater, and an unpleasant odor spreads through the surrounding area. In addition to that, waste products start burning once they come into contact with air, which generates fires that are very difficult to extinguish, carrying pungent smoke into residential areas.
In the course of the experiment, we adhered to the following hypothesis: the drone allows us to check the test site and identify possible violations effectively and quickly.
According to the legend of the experiment, an alarm signal is sent to the base, which is located on the territory of the technopark, about a possible violation at the landfill. Automatically the drone takes off from the base and is sent to the known coordinates to the landfill. Arriving at the destination, the drone begins circling the area, fixing everything onto a 4K camera. After completing the flight, the drone returns to the base. And on the base the operator on duty uploads a video, watches it and identifies violations.
In terms of software, everything looks like this:
- When an alarm signal is received at the base, the demand for exploring the landfill, which is sent to the markets of Robonomics, is formed.
- Once the system finds the drone available for fulfilling the task, the lighthouse of the Robonomics network matches the supply and demand. Mission for research is assigned to this drone.
- A smart contract is then formed for this particular drone, which prescribes target coordinates, the need for video footage and other necessary information.
- The flight to the mission is in progress.
- After the mission is accomplished, the drone automatically returns to the base and transfers the link to the distributed IPFS file system where the video is stored. Now the video is available for viewing.
The experiment was conducted on August 7, 2018. The drone flew over to the landfill after having received the alarm signal, successfully completed the video shooting and returned safely back to the base. The processing time for departure was 1–5 minutes. The time to move to the target was 1 minute 50 seconds. Time of exploration of the landfill was 1 minute 20 seconds. The total time of the assignment completion was about 10 minutes.
For comparison: the trip of the operational services brigade to the landfill from the moment of receiving the alarm usually takes 20–30 minutes. Because of this, it is impossible to legally record the fact of the offense, and violators have time to escape. And thanks to the drone, processing of the signal is very fast and goes unnoticed for violators.
The experiment over the forest area
Venue of the experiment: health resort “Lesnoye”, Kuneyevsky forest, Togliatti, Russia
The districts of Togliatti are separated from each other by large forest areas, whose total area is almost equal to the area of the urban part. Such a significant forest area requires monitoring of fire safety, monitoring of suspicious activity in the forest area and control over illegal logging. Traditional methods of monitoring (ground patrolling, observation towers, inspection from a helicopter) either take a long time, or require high costs. We decided to check, if it is possible to provide high-quality and easily accessible monitoring of forest areas with the help of an UAV.
The experiment is based on a similar legend. On the base located on the territory of the health resort “Lesnoye” comes an alarm signal about a possible violation in the forest zone. Together with the signal coordinates are received. From the base, the drone automatically flies up to the mission and starts flying to the received location, and the video camera on the drone starts an online broadcast. At the same time, an alarm notification is sent to the attendant’s computer and a pop-up window opens where the attendant can watch the broadcast. Having reached the place of violation, the drone hangs over it, trying to capture the territory as much as possible. The average observation time is 30 seconds. Based on the situation, the duty officer at the base takes a decision on emergency measures. After completing the surveillance, the camera stops broadcasting, and the drone automatically returns to the base.
And here you can see the software work for the drone mission:
- When the alarm signal and the coordinates of the incident are received at the base, the demand for the mission to the given location is formed. Demand is published in the markets of Robonomics.
- The lighthouse of the Robonomics network finds an offer that is corresponding to the demand.
- A smart contract is made with all the necessary mission data for the selected drone.
- Drone starts the mission.
- During the mission, the drone publishes an IPFS link to the broadcast on YouTube. The operator gets access to the surveillance.
- After observing, the drone automatically returns to the base, and informs the network about the completed mission.
The experiment was conducted on August 10, 2018. Drone successfully completed the flight to the given coordinates, the video shooting was successfully launched and broadcast for 4 minutes 50 seconds. The processing time for the order was 1–5 minutes. The flight time to the given coordinates was 2 minutes 50 seconds, the distance to the base from the target was 1.25 km.
For comparison: the time to reach the location by car on unsurfaced roads is 6 minutes, but not all areas of the forest are available for travel.
As we can see with the help of the drone, monitoring of hard-to-reach territories is simplified and becomes more efficient. Even one UAV of a popular model is capable of performing automatic monitoring of 5 km2 territory around the base with a return back, spending for the entire mission up to 7 minutes. For the model of the drone used in the experiments (DJI Matrice 100, cost $ 3800) with a working load in the form of a camera, the operating time without recharging is about 30 minutes, which gives us 4 consecutive missions. For comparison: helicopter flight costs from $ 730 per hour, depending on the brand. This shows how much the use of drones for observation is advantageously different from traditional monitoring methods. The drone mission is simpler in organization and cheaper in cost than the helicopter’s mission, and ground surveillance and observation towers do not have the mobility and efficiency inherent in drones.