WO2025088161A1

WO2025088161A1 - A risk analysis based system for infrastructure protection

Info

Publication number: WO2025088161A1
Application number: PCT/EP2024/080297
Authority: WO
Inventors: Dawid GRADOLEWSKI; Adam Jaworski; Damian DZIAK; Damian KANIECKY; Wlodek KUSLESZA
Original assignee: Bioseco SA
Current assignee: Bioseco SA
Priority date: 2023-10-26
Filing date: 2024-10-25
Publication date: 2025-05-01
Anticipated expiration: 2026-04-26

Abstract

According to the invention there is method for assessing the strike risk probability of a flying object and a protected infrastructure, the method comprising detecting a position of the flying object at a successive points in time, predicting a path based on the on the plurality of detected positions, determining a position risk factor based on the current position related to the protected infrastructure and determining an overall strike risk based on a combination of two factors the path risk factor and the position risk factor.

Description

A RISK ANALYSIS BASED SYSTEM FOR INFRASTRUCTURE PROTECTION

The invention relates to a system and method for mitigating the risk of collisions of a flying object with a protected infrastructure. Examples of such infrastructure are wind farms and airports, which are exposed to a risk of collision with flying objects such as avifauna and drones.

Collisions of objects such as animals and drones with wind turbines and other high value installations can lead to accidents, which are endanger wildlife and/or people. Additionally, damages to infrastructure may also be caused, which can increase maintenance and repair costs.

Mono- and stereo-vision, radars and/or lidar sensors can be used to detect flying objects and hazard situations. Detection systems based on such sensors are often used in higher risk areas such as the airspace of wind farms and airports. Once a flying object is detected in a risk area, the relevant avoidance measures are activated.

However, the demanded range of detection systems is often extensive because the time taken to stop a turbine can be significant. This can result in a high number of false positive alarms when a flying object is within the detection range but does not pose a real collision threat. Also, the use of unnecessary repellent measures caused by a false positive alarm can cause wildlife stress and leads to habitation to repellents. Stopping the wind turbine results in costly downtime for the apparatus as a whole.

It is therefore an object of the invention to provide and improved a method of reliably assessing in real time the risk of a collision.

According to the invention there is provided a method for assessing the collision risk probability of an object and a protected infrastructure. The method comprises:

• determining positions of the object at successive points in time and estimating a position risk factor based on the current position of the detected object and a protected infrastructure collision area;

• predicting a path based on the detected successive positions in time, and estimating a path risk factor based on the predicted path and the prediction uncertainty represented by a solid angle; • and determining an overall risk factor based on a combination of the position risk factor and the path risk factor.

In one embodiment the prediction uncertainty includes the predicted path.

Advantageously, the invention provides a method of assessing the risk which takes into account both the location of the protected infrastructure and also a predicted path. It provides a method of quantifying the risk of a collision.

The method may be used in particular with stereovision systems.

The position risk level can be visualised as a cone from the current position of the object to the plane covering the potential collision risk area. The path risk level can be visualised as a cone from the current position of the object to plane of the potential collision risk area, the central axis of the cone is the predicted path, and the angle of the cone indicates the uncertainty. The overall risk level can be illustrated as the overlap of these cones.

The position risk factor may be estimated based on a solid angle defined by the current position of the object and the protected infrastructure collision area.

The position risk factor may be based on the minimum distance for collision avoidance TT. In particular, the minimum distance for collision avoidance is given by r? = Vmax X T_stop, wherein Vmax is the maximum speed of the flying object and T_sto_P is the minimum stopping time of the moving parts of the protected infrastructure. n r-r

The position risk factor may be given by P_pos = - X — wherein r is the

^max ^r distance of the flying object to the closest point of the protected infrastructure, Q is the solid angle defined on the relation of the protected infrastructure collision area and the current position of the flying object and Qmax is the maximum possible solid angle defined for the worst case position related to the protected infrastructure collision area.

Optionally, the position risk factor is related to an object coefficient based on the type of detected flying object in a given environment. The path risk factor may be based on a solid angle defining the future path space based on trajectory prediction and prediction uncertainty. In particular, the path risk factor may be given by:

where H_coii is a binary coefficient based on whether the solid angle predicted path space intersects with the surface of the protected infrastructure; Qcrmax is a maximum solid angle for the worst case of confidence level of the estimate of tracking risk space and v|j_con is a collision angle between the predicted collision direction and a plane of protected infrastructure.

The angle v|j_con can be determined from:

where AP_n and P_nP_n+1 are vectors defined by the three points A, P_n, P_n+1.

According to the invention there is provided a risk assessment system comprising a processor configured to carry out the steps as described above.

The system may further comprise an object detection system configured to detect objects and transmit data about the detected object to the risk assessment system. A localisation and identification system is configured to identify and localised position of an object detected by the object detection system and transmit data about the object’s identity and localisation to the risk assessment system. A tracking system configured to anticipate the position of the detected object from successive positions and transmit data about the predicted path of the detected object to the risk assessment system. A decision making system is configured to make a decision based on the determined overall risk; and activating a collision avoidance system.

Optionally the system further comprises one or more environmental sensors configured to detect environmental factors and configured to transmit data about the environmental factors to the risk assessment system.

Figure 1 depicts a block diagram of the system according to the invention; Figure 2 illustrates a top view of the collision situation with a view of cross section of the cons illustrating risk areas due to the object’s position and the tracking;

Figure 3 depicts top view of the collision situation illustrating a risk area due to the object’s position including all related quantities;

Figure 4 depicts a front view of the collision situation illustrating the position of the flying object relative to the collision area of a fixed object, including all related quantities;

Figure 5 depicts the detected path of a flying object and a cross-section of the cone depicting the predicted path.

Figure 1 depicts a block diagram of a proposed system in which an object detection unit detects objects using, for example, one or more of thermal images, infrared, vision, and/or stereo vision sensors. These are used to detect a moving object in the monitoring area. The localisation and identification unit identifies the type of flying object and also its position in 3D space. An object tracking unit depicts the object’s trajectory and also applies an algorithm to predict the future movement of the flying object. Information from the detection unit, localisation and identification unit and object tracking unit is provided to a risk assessment unit.

An environmental sensors unit comprises such devices as anemometers, wind direction indicator, visibility and humidity sensor, rain sensor, temperature and/or pressure sensor. The unit provides information about environmental conditions which can impact detection process and or objects path characteristics. Data from the environmental sensors is transmitted into the risk assessment system. The risk assessment system continually assesses the collision risk level, and the operation of the risk assessment system is described in more detail below. The decision-making system interacts with the risk assessment system and the collision avoidance system to control the apparatus to avoid collisions.

The object risk probability associated with a moving object colliding with an infrastructure, P_Obj is a product of two factors: P_pos, the position risk factor dependent on the current positions and Ptrack the path risk factor associated with the anticipated path of the flying object. IPobj- IPpos X Ptrack

These risk probabilities are affected by the uncertainty of any measurements on which they are based on, the reaction time of the system(s) and the size of the risk area. The risk area/space associated with the position can be visualised as a solid angle with an apex at the current position of the flying object and covering a possible collision area of the protected infrastructure. The risk area/space associated with the tracking can be visualised as a solid angle/cone, the central axis of which is the predicted path and the angle of the cone is defined by the prediction uncertainty. These two collision risk areas/spaces are depicted in Figure 2 in which a first cone depicts the position risk space for a flying object colliding with the protected infrastructure. A second cone depicts the risk space related the predicted flight path. The intersection of these two spaces depicts risk level of the flying object colliding with the protected infrastructure. In the example depicted in Figure 2 there is a low risk of collision as there is no intersection between the two risk spaces.

In Figure 3, a shortest distance for collision avoidance PT is depicted. Based on a maximum known speed of the detected flying object, Vmax together with the longest stopping time for the turbine, T_sto_P, a shortest distance needed to activate collision avoidance system can be calculated:

TT ⁼ Vmax X Tstop

Tstop may comprise both the reaction time and the latency time of the system. Based on the detected type of object, or species of bird, the known maximum speed for that particular species can be used.

The risk space associated with the position of the flying object is defined as the solid angle Q from a flying object position P_n covering the space of possible collision of radius R. The radius R is the radius over which the flying objects could have direct contact with the protected infrastructure. And r is the shortest distance between an object position P_n(x_n, y_n, z_n) and the closest point of the fixed object A (x_a, y_a, 0) and this is depicted in Figure 4. The furthest point of the protected infrastructure is position B(xb, yb, 0) where x_a= -Xb and y_a= -yb.

The position risk factor Pp_0S(Q, r, t) can be estimated using:

where:

- Q is a solid angle (in steradian) corresponding to a potential collision space defined by segments connecting the object’s position point P_n(x_n, y_n, z_n) with the closest point of the protected infrastructure A(x_a, y_a, 0) and the furthest point of the fixed object B(Xb, yb, 0);

- Qmax is a solid angle (in steradian) corresponding to the maximum possible collision space when the detected object's trajectory is perpendicular to the centre of collision surface at a distance TT.

- S is the surface area of a base of the cone defined by angle Q and the distance r between the flying object’s position P_n(x_n, y_n, z_n) and its closest point on the fixed object A(x_a, y_a, 0);

- r is the distance between the flying object and the closest point of the protected infrastructure;

- 2R is the diameter of the protected infrastructure collision area.

The angle 20 is an apex angle: Q=2 IT /(1 -COS 0) where

where AP_n and P_nB are vectors defined by the three points A, P_n, B.

The angle 20_max is an apex angle: Q_max=2 IT /(1 -COS 0_max) defined per analogiam to the apex angle 20 and is depicted in figure 3.

Then the equation defining the position risk factor P_pos(Q, r, t) could be simplified as:

The tracking risk factor Ptrack is based on previously detected positions Pn-i, Pn-2, Pn-3.... This is depicted in Figure 5. The future flight trajectory may be predicted using a tracking algorithm such as a Kalman filter, Multiple Hypothesis Tracking or other methods. The predicted flight trajectory forms the axis of a cone and the uncertainty defines a cone around the axis and these are depicted in Figure 5. The tracking risk factor can then be calculated as:

where:

- Hcoii is a binary coefficient (0 or 1 ) based on whether the tracking path intersects with the protected infrastructure;

- ijicon is ^a collision angle between the predicted collision direction and a plane of protected infrastructure;

- is the solid angle representing tracking uncertainty;

- Qcrmax is the maximum solid angle for the worst case accuracy tracking for a particular site/process. This represents the prediction worst case uncertainty for the highest confidence level. Qcrmax is estimated from the statistical properties of the prediction process.

As an example, if

is estimated as a variance of the distribution of N points in 3D space, then may be a multiple

. It could be a 3o or 4o confidence level.

The equation above can be simplified to:

where:

In addition to the variables described above which affect the position risk factor and the tracking risk factor there may also be a heuristic object coefficient H_Obj which is determined for a particular situation. The value of H_Obj is specific to the particular type of flying object and the particular site i.e. a different species of bird may have a different Hobj and the same species of bird at a different site may have a different Hobj. Hobj is therefore customised for different sites and may be amended over time based on gathered data. Using the heuristic coefficient, the factors P_pos and Ptrack would be:

Strack = H_obj x cosv|j_COH I

_ 1 — cosSg.

+ H_C0u _ _COSQ

°max

The object collision risk probability Pobject related the object position and its trajectory has been discussed above. However, environmental conditions such as wind speed and direction can impact birds’ behaviour. The system performance may be affected due to visibility, humidity, rain, temperature, air pressure and other factors. There is also a risk factor associated with environmental factors such as the wind speed and direction. The overall strike risk probability Pstrike could be described as:

P strike⁼aPobjecC*"[3Penvironment where a and [3 are weighting coefficients such that a + (3=1 . The coefficients a and [3 define the impacts of the object and the environment components on overall collision risk probability and are site specific. These coefficients can be determined heuristically and may be adjusted over time using, for example, a self-learning algorithm.

Although the invention has been described in relation to wind farms it can equally be used at an airfield or by an aircraft.

Various further aspects and embodiments of the present invention will be apparent to those skilled in the art in view of the present disclosure.

“and/or” where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example, “A and/or B” is to be taken as specific disclosure of each of (i) A, (ii) B and (iii) A and B, just as if each is set out individually herein. Unless context dictates otherwise, the descriptions and definitions of the features set out above are not limited to any particular aspect or embodiment of the invention and apply equally to all aspects and embodiments which are described.

It will further be appreciated by those skilled in the art that although the invention has been described by way of example with reference to several embodiments. It is not limited to the disclosed embodiments and that alternative embodiments could be constructed without departing from the scope of the invention as defined in the appended claims.

Claims

1 . A method for assessing the collision risk probability of an object and a protected infrastructure, the method comprising: detecting a position of the object at successive points in time; determining a position risk factor based on the current position of the detected object and a protected infrastructure collision area; predicting a path based on the detected successive positions in time; determining a path risk factor based on the predicted path; and the prediction uncertainty represented by a solid angle; determining an overall risk factor based on a combination of the position risk factor and the path risk factor.

2. A method according to the preceding claims wherein the position risk factor is estimated based on a solid angle defined by the current position of the object and the protected infrastructure collision area.

3. A method according to claim 2 wherein the position risk factor is based on the ratio of minimum distance for collision avoidance r?and an object distance from the protected infrastructure.

4. A method according to claim 3 wherein the protected infrastructure comprises moving parts and the minimum distance for collision avoidance is given by TT = Vmax X Tstop, wherein Vmax is the maximum possible speed of the flying object and Tstop is the minimum stopping time of the moving parts of the protected infrastructure.

5. A method according to claim 4 wherein the position risk factor given by wherein r is the distance of the flying object to the closest

point of the protected infrastructure, Q is the solid angle defined on the relation of the protected infrastructure collision area and the current position of the flying object and Qmax is the maximum possible solid angle defined for the worst case position related to the protected infrastructure collision area.

6. A method according to any one of the preceding claims wherein the position risk factor is based on an object coefficient based on the type of detected flying object in a given environment.

7. A method according to any one of the preceding claims wherein the path risk factor is based on a solid angle defining the future path space based on trajectory prediction from trajectory history and prediction uncertainty.

8. A method according to claim 7 wherein the path risk factor is given by:

where v|j_con is a collision angle between the predicted collision direction and a plane of protected infrastructure; H_coii is a binary coefficient based on whether the solid angle predicted path space intersects with the surface of the protected infrastructure;

is the solid angle defining the predicted path space; Qcrmax is a maximum solid angle for the worst case of confidence level of the estimate of tracking risk space.

9. A method according to either claim 7or claim 8 wherein the path risk factor is based on an object coefficient based on the features of detected flying object in a given environment.

10. A method according to any one of the preceding claims wherein the overall object risk factor, P_Obj, is a product of the position risk factor and the path risk factor:

Pobj⁼Ppos X Ptrack

11. A method according to any one of the preceding claims wherein the overall strike risk factor is based on environmental factors.

12. A method according to any one of the preceding claims wherein the overall strike risk probability Pstrike could be described as:

P strike⁼C(Pobject+[3Penvironment where a and [3 are site specific weighting coefficients such that a + (3=1 and a and [3 define the impacts of the object and the environment components respectively on overall collision risk probability.

13. A risk assessment system comprising a processor configured to carry out the steps according to any one of the preceding claims.

14. A system according to claim 13 further comprising: an object detection system configured to detect objects and transmit data about the detected object and its position to the risk assessment system; an identification system configured to identify an object detected by the object detection system and transmit data about the identified object to the risk assessment system; a tracking system configured to track the position of the detected object based on successive positions and transmit data about the predicted path of the detected object to the risk assessment system; a decision making system configured to make a decision based on the determined overall risk; and a collision avoidance system configured to activate collision avoidance measures.

15. A system according to claim 14 further comprising one or more environmental sensors configured to detect environmental factors and configured to transmit data about the environmental factors to the risk assessment system.