*2.2. Radar Used in this Study*

The specifications of the radar used in this study are presented in Table 1. It is a 3D UMRR 42HD automotive radar with a 24 GHz microwave sensor. The type 42 antenna has a wide field of view. The sensor is a 24 GHz 3D/UHD radar for motion management and is able to operate under adverse conditions, measuring in parallel parameters such as angle, radial speed, range, and reflectivity. It is usually used as a standalone radar for detecting approaching and receding motion. More details on the sensor used can be found in [23].


**Table 1.** Specifications of the UMRR 0C Type 42 anti-collision radar.

#### **3. Sensors for ASV Autonomous Anti-collision**

As it was said, the anti-collision systems in ASVs are part of the wider concept which aims to provide tools for the safe and reliable navigation of vessels. One of the key elements of such a systems are the sensors, which provide data about the environment for further processing.

#### *3.1. Situation Awarness Systems for ASV*

Previous research has addressed the various aspects of navigation of unmanned vehicles. Video data and LiDAR fusion are described in [24]. In [25], an algorithm using LiDAR and a camera for detecting and tracking surface obstacles using the Kalman filter is presented. The legal aspects of ASV navigation, including anti-collision, are described in [26]. An approach using artificial neural networks (ANN) to solve the ASV anti-collision problem is presented in [27–29], where ANN was used to control the autonomous robot. The 3D mobile (3D LiDAR) and GNSS applied to autonomous car navigation was presented in [30].

One of a few attempts to use both radar and LiDAR in the navigation of mobile robots was described in [31–33], which also highlighted new development directions for land mapping based on radar and LiDAR. An approach using radar and LiDAR fusion to detect obstacles was taken in [34], an attempt to replace the radar with LiDAR was shown in [35] and the aspects of obstacle sensing by synthetic aperture radar interferometry was presented in [36]. An interesting approach of the anti-collision system for ASV is described in [37] in which the gathering of situational awareness relies on GPS and AIS.

Target detection in close range observation is very important to provide the next step in ASV navigation, which can be achieved by developing an advanced fast filter to track targets at a close range using automotive 3D radar. Neural solutions for radar target tracking by maritime navigation radar have been previously described by the authors of this study [38,39].

#### *3.2. Anticollision Based on Radar Systems*

As it was mentioned in the introduction, there are also several examples in literature of using radar target detection and tracking for the anti-collision of ASV. These examples can be found for example in [9–11], but also in [40] in which the anti-collision system based on the sensors traditionally used on maritime ships is presented.

A radar sensor is a commonly used device for anti-collision at sea. Two kinds of solutions are used for the marine environment—X-band radars and S-band radars. Both of them have their advantages and disadvantages, however from the ASV point of view none of them are suitable. The reason for this is that they both require relatively large antennas to achieve reasonable resolution. Such antennas (of a few meters wide) cannot be mounted on the ASV, which are usually floating platforms of a few meters in length. On the other hand, the advantages of the radar technology and its usefulness for anti-collision purposes are hard to be overestimated. Radar waves are relatively resistant to environmental clutters and thus can be used in fog or even rainy conditions in which cameras and lidars are useless. Taking all this into account, as well as experience in radar data processing, we were looking for the possibilities of providing radar technology for anti-collision purposes in ASVs. Thus, an idea arose to use radars used in automotive applications for tracking and anti-collision in HydroDron. The main advantages of this proposed, novel solution are:


The innovative approach is however as always burdened with some risk. The important questions are—how will the 77 GHz radar deal with a water environment, how will surface targets be detected in this type of radar, what will the detection ranges be in this implementation, and what particular processing techniques will be useful for this radar in this particular implementation. The answers for these questions have to be found and for this reason suitable research is needed. In this study, the goal was to empirically verify the detection possibilities using automotive 3D radar in the water environment, as most previous research using these types of radar systems have been conducted in onshore conditions. The radar observations of various targets on the water were collected using a radar mounted on an ASV.
