Page 64 - Read Online

Page 4 of 15 Ma et al. Intell Robot 2023;3(4):581-95 I http://dx.doi.org/10.20517/ir.2023.33

Figure 1. Structure of task definition.

2. PROBLEM FORMULATION
In modern intelligent systems, the heterogeneous multi-agent assignment problem is a research field of great
concern. This problem involves various types of intelligent agents, such as drones, unmanned ships, and un-
manned vehicles, deployed in unknown environments to perform various tasks. These intelligent agents have
different sensors, actuators, motion capabilities, and various resource constraints. In complex, dynamic, and
unfamiliar environments, such as exploring unknown areas or seas, these agents need to collaborate effectively
to accomplish tasks.

2.1. Description of heterogeneous multi-agent models
The heterogeneous multi-agent model is a research model that involves task allocation and collaboration issues
for different types of agents, such as drones, unmanned ships, and unmanned vehicles. In this model, different
types of agents have varying task execution capabilities, resource allocations, and perception abilities, and they
need to cooperate to efficiently complete a complex set of emergency rescue tasks.

Definition of intelligent agent types:

• Drone: Equipped with high-resolution cameras and sensors, capable of swift flying to obtain images and
data of disaster areas. It can also search for personnel and survey dangerous areas.
• Unmanned vessel: Equipped with diving sensors and life-saving equipment, it can carry out search and
rescue missions in the water, rescuing trapped personnel and providing emergency rescue supplies.
• Unmanned vehicle: With strong off-road capability and the ability to carry materials, it can enter affected
areas for personnel evacuation, material transportation, and medical support.

Task Definitions:

• Disaster investigation task: After a disaster occurs, drones are dispatched to the affected area for rapid
investigation. They can capture high-resolution images, monitor disaster severity, and provide real-time
data to the command center to help formulate rescue plans.
• Personnel search and rescue task: Unmanned ships and vehicles are responsible for searching for trapped
personnel in the affected waters and on the ground. Unmanned ships can detect underwater conditions
through diving sensors, while unmanned vehicles can search for buried personnel using sensors.
• Material transportation task: Unmanned vehicles are responsible for transporting emergency rescue mate-
rials from the command center to the disaster area, providing disaster relief materials, and offering medical
support.

Communication and collaboration mechanism: Establish a high-speed wireless communication network be-
tween UAVs, unmanned ships, and unmanned vehicles, using satellite communication or dedicated frequency
bands to enable real-time data and status information sharing. By employing distributed decision algorithms,
intelligent agents can exchange task information and sensor data in real time, enabling them to determine task
allocation and execution strategies through consensus algorithms.

Thedisasterinvestigationmissioniscarriedoutbydronesequippedwithfastflightandhigh-resolutionsensors