Transport). In the first case, drones can be used to deliver goods in different areas with the advantage of saving

costs and delivery times; in the second case, drones can be used to monitor and control the flow of people

inside public transport, to detect any situations of overcrowding and take prompt action to ensure the safety and

comfort of passengers. This research aims to propose an innovative automatic management system for a group

of RPAS used for the transport of essential medicines and for monitoring the crowding of people in relation to

the use of Local Public Transport. This was possible through the creation of an experimental system for

recharging drone batteries and above all the implementation of different soft computing algorithms and tools

such as YOLO (You Look Only Once), SORT, and MAC address. All the information obtained in real-time and

continuously updated, are transmitted to an Open GIS studied and programmed by us.

different fields due to their ability to fly and collect

aerial data quickly and efficiently. Nowadays,

research is progressively directed towards the

automated flight of routes, to make the use of drones

more efficient and safer. In addition, the research is

focusing on the extent of the flight duration and the

use of different sensors, so that the full potential of

drones can be exploited. In the context of delivery

and crowd monitoring, in the literature, many

applications can be found, [1], [2], such as in the

other dedicated methodologies and technologies. To

overcome these limitations, our research was

focusing on the use of GIS (Geographic Information

System) and different soft computing algorithms to

give an important contribution beyond the actual

studies present in the literature both in relation to

photogrammetry using drones in the evaluation of

monitoring systems that use artificial vision

techniques to identify vehicles, pedestrians, and

obstacles on the road, [13], [14]. In this context, we

presented a system that allows, at the same time,

study area only permit VLOS (Visual Line of Sight)

operations, our fleet of drones is fitted with safety

parachutes that are triggered by an accelerometer.

This parachute is compatible with drones that have

sensors capable of detecting flight interruptions. The

sensor sends a command to an actuator that deploys

the parachute, which is protected from the propellers

to prevent damage.

drones, a self-implemented GIS, and a series of

different soft computing algorithms and tools.

Among all the drones available, we have chosen the

DJI Matrice 100 (Fig. 1) due to its technical

characteristics, which were useful for the study we

conducted. This type of drone is furnished with

omnidirectional vision sensors, an obstacle detection

system, intelligent features such as Point of Interest,

allowing the drone to keep easily the point centered

in the camera creating complex shoots, and are

autonomy. DJI Matrice 100 has a 4500 mAh TB47D

6S battery with an autonomy of some minutes that

we have managed to extend with a wireless charging

system described below.

raise problems concerning the privacy of public

transport service users as the YOLO algorithm does

not identify a specific person. Tracing, on the other

hand, is possible thanks to the SORT algorithm and

the MAC address which is immediately obfuscated

through the application of a cryptographic function.

An origin/destination matrix is thus constructed and

updated on an hourly basis. The aim was to follow

people along the path from the origin to destination

using Machine Learning techniques and to do that

The flow present in each cell of the matrix was

subdivided into the numbers of users using a certain

line of the LPT, also using the MAC address. We

have created, therefore, a system to obtain complete

O/D matrices automatically in a GIS programmed in

Python. The aim was to generate hourly forecasts of

flow data and continuously monitor trends by

comparing real-time data with historical data.

composed of distinct phases (both of which use the

of the elaboration of the images acquired using ad

hoc algorithms and integrating the results with other

algorithms to define the origin-destination matrix

and to identify the movements of people. In relation

to the methodology used for the delivery of

necessities and the monitoring of crowding of

people, the flow chart in Fig. 9 shows the various

stages of the process which involved the algorithms

described above.

the city, therefore a sample area; in fact, the purpose

is only to show the proposed research method, but it

can also be implemented on wider areas of interest.

We have analysed step by step the procedure for

delivering goods and monitoring the crowd.

Fig. 10: Study area, a neighborhood of Reggio

In this area we have established:

Fig. 11: GIS visualization of the medical goods’

delivery.

As can be seen from Fig. 14, the thickest lines are in

Via Marina which connects the center to the

outskirts of the city and where most of the public

(Fig. 13) the degree of crowding through circles that

grows as the number of people waiting at the bus

stop increases.

bus or to distribute the LPT users.

information on the crowding of people. All thanks

also to the innovative recharging system that allows

you to extend the autonomy of the drone without the

need for personnel present on site. Users will thus

be able to choose whether to take one line rather

than another and the institutions will be able to

better-performing hardware and software. Through

the proposed method, it was possible to deliver

small goods and monitor the degree of crowding at

bus stops, but there are several and many future

developments of this technology that uses soft

computing and machine learning techniques. Some

of the possible future directions in this field include

real-time data analysis that would enable the

identification of patterns and trends in the data, as

well as the prediction of future crowd movements;

waiting at bus stops or available seats on

transportation. Providing this information is

important for public transportation for several

reasons. They allow for improving passenger

experience by reducing the stress and anxiety of

companies to better plan bus routes.