Collision avoidance device, collision avoidance method, and program

Abstract

This collision avoidance device comprises: a passing-gap estimation unit that estimates a passing gap between the host vehicle and an avoidance target on the basis of a relative speed vector and a relative position vector between the host vehicle and the avoidance target, the avoidance target being a passing vehicle, from among passing vehicles, of which the relative distance to the host vehicle is decreasing temporally; and a speed vector determination unit that determines, on the basis of the passing gap, a speed vector for causing the host vehicle to perform a collision avoidance action.

Description

Collision avoidance device, collision avoidance method, and program

[0001] This application claims priority to Japanese Patent Application No. 2023-123124, filed on July 28, 2023, the contents of which are incorporated herein by reference.

[0002] Collision avoidance technology is important for simultaneously navigating multiple unmanned aerial vehicles (unmanned aerial vehicles). A common collision avoidance technology involves, for example, having a person constantly remotely monitor the separation between unmanned aerial vehicles from the ground. If the separation between certain unmanned aerial vehicles becomes abnormally small, the person on the ground issues a command to at least one or more of the unmanned aerial vehicles to change course or speed to avoid a collision. However, if wireless communication between the ground and the unmanned aerial vehicles fails, this method of issuing commands from the ground will not work.

[0003] Another example of a common collision avoidance technology is the Traffic Alert and Collision Avoidance System (TCAS) for piloted aircraft (manned air vehicles). TCAS monitors the aircraft's surroundings independently of ground-based air traffic control systems, and if there is another aircraft that could potentially cause a collision, it provides the pilot with an audio collision warning and instructions for avoidance. However, TCAS is not effective for unmanned air vehicles.

[0004] Therefore, there is a need for a technology that automatically avoids collisions between vehicles, including unmanned vehicles. For example, Patent Document 1 (Patent Document 1) describes a technology that uses line segments and arcs to represent an avoidance trajectory for an airborne vehicle to avoid a collision with another airborne vehicle (loss of separation). The parameters required to determine the line segments and arcs (e.g., the duration or angle value of the turn represented by the arc) must be determined using an iterative calculation such as the Newton-Raphson algorithm. In other words, the avoidance trajectory cannot be determined using a simple one-way function. This results in a high computational load. Furthermore, in the method of determining the avoidance trajectory, the number of line segments and arcs representing the avoidance trajectory increases when a new target for avoidance appears during the avoidance process, further increasing the computational load.

[0005] Furthermore, Patent Document 2 describes using an avoidance command that changes the vehicle's (aircraft's) heading, climb rate, descent rate, speed, etc., rather than determining an avoidance trajectory using a line segment or an arc. Because the avoidance command can be determined using a one-way function, the calculation load can be reduced compared to Patent Document 1.

[0006] Japanese Patent No. 5933244 Japanese Patent No. 6272637

[0007] However, with the conventional technology exemplified by Patent Document 2, it may be difficult to avoid a collision when multiple vehicles are passing by.

[0008] A specific example will be described. In Patent Document 2, when a first aircraft, which is the subject vehicle, and a second aircraft, which is another vehicle, fly within a certain airspace, in order to manage the distance between the first aircraft and the second aircraft, an area where both the first aircraft and the second aircraft must not be present at the same time is defined by a virtual boundary line (separation boundary). Furthermore, a collision avoidance module of the first aircraft predicts the closest point of approach between the first aircraft and the second aircraft, and if the range between the first aircraft and the second aircraft at the closest point of approach violates the separation boundary, an avoidance command is issued to change the flight path of the first aircraft. Note that the collision avoidance system calculates the required level of separation m ｒｅｑｕｉｒｅｄ , safety level interval m ｓａｆｅ , or the interval m of the specified level ｉｍｐｏｓｅｄSeparation perimeter layers may be defined in three dimensions.

[0009] The collision prevention system determines a collision prevention command component C- for each separation boundary. If there are multiple aircraft passing other than the own aircraft, the collision prevention system determines an avoidance command for each of them other than the own aircraft. Therefore, if the number of separation boundaries is p and the number of passing aircraft is n, the number of collision prevention command components will be n x p. For simplicity, the explanation will be given assuming the number of separation boundaries is 1. In this case, the collision prevention command component is {C- １，１ , C- １，２ , ..., C- １，ｎ} and there are n collision prevention command components. Even if the number of aircraft passing through the avoidance command is n, there is only one each of {heading, climb rate, descent rate, speed} of the own aircraft, so it is not possible to control according to multiple collision prevention command components. One {heading, climb rate, descent rate, speed} of the own aircraft must be determined from multiple avoidance commands. In Patent Document 2, the collision prevention system uses a gain K ｐ，ｎ The avoidance command F- is calculated by weighting the collision prevention command components. ｃｏｎｔｒｏｌ When the number of interval boundaries p is 1, the avoidance command F- ｃｏｎｔｒｏｌ Make a decision.

[0010]

[0011] However, the weighted sum is not always appropriate for the purpose of collision prevention. For example, consider the situation shown in Fig. 12. Fig. 12 shows an example in which the own aircraft 90 intersects with the passing aircraft 91 and the passing aircraft 92. At time t ０ The positions of the aircraft 90, 91, and 92 in the solid line triangle P90 ０ , P91 ０ , P92 ０ If each of the aircraft 90, 91, and 92 maintains the original course 900, 910, and 920, the time t ｘ The positions of the aircraft 90, 91, and 92 are indicated by dashed triangles at position P90.ｘ , P91 ｘ , P92 ｘ If the passing aircraft 91 and the passing aircraft 92 maintain their original routes 910 and 920, the collision avoidance command component C- １，１ Change of course to the left of the page (θ １ The changed course 901 is suitable. ｘ The aircraft 90 is at position P90 ｘ＿１ At this time, the distance between the own aircraft 90 and the passing aircraft 91 is d, which is a value that does not violate the distance boundaries B90 and B91. Similarly, the collision prevention command component C- １，２ Change of course to the right of the page (θ ２ The changed course 902) is appropriate.

[0012] At this time, the collision prevention command component C- １，１ and collision avoidance command component C- １，２ , the gain K １，１ = 0.5 and gain K １，２ = 0.5 and according to the formula (1), the course change to the left and the course change to the right cancel each other out, and the original course 900 is maintained as the avoidance command F- ｃｏｎｔｒｏｌ This makes it difficult to avoid violating the boundary of the separation.

[0013] In the example of FIG. 12, the distance d between the own aircraft 90 and the passing aircraft 91, 92 is roughly as shown in FIG. 13 with respect to the course change angle θ from the original course 900. １，１ (Left course change θ １ ) and Collision Prevention Command Component C- １，２ (Changing course to the right θ ２ ) as in equation (1), the "intrusion" cannot be avoided. In the situation of FIG. 12, the interval d is within the range [θ １ , θ ２] is not a convex function, so it is not easy to avoid "violation" using numerical search methods.

[0014] An object of the present disclosure is to provide a collision avoidance device, a collision avoidance method, and a program that can avoid a collision even when multiple vehicles are passing by.

[0015] According to one aspect of the present disclosure, a collision avoidance device is a collision avoidance device that avoids a collision between a vehicle and another passing vehicle, and targets a passing vehicle among the passing vehicles whose relative distance to the vehicle decreases over time as a target to be avoided, and is equipped with a passing interval estimation unit that estimates a passing interval between the vehicle and the target to be avoided based on a relative speed vector and a relative position vector between the vehicle and the target to be avoided, and a speed vector determination unit that determines a speed vector that causes the vehicle to take collision avoidance action based on the passing interval.

[0016] According to one aspect of the present disclosure, a collision avoidance method is a collision avoidance method for avoiding a collision between a vehicle and another passing vehicle, and includes the steps of: selecting a passing vehicle among the passing vehicles whose relative distance to the vehicle decreases over time as an avoidance target; estimating a passing interval between the vehicle and the target to be avoided based on a relative velocity vector and a relative position vector between the vehicle and the target to be avoided; and determining a velocity vector that causes the vehicle to take collision avoidance action based on the passing interval.

[0017] According to one aspect of the present disclosure, a program causes a collision avoidance device that prevents a vehicle from colliding with another passing vehicle to perform the following steps: identify a passing vehicle among the passing vehicles whose relative distance to the vehicle decreases over time as a target to be avoided; estimate a passing interval between the vehicle and the target to be avoided based on a relative speed vector and a relative position vector between the vehicle and the target to be avoided; and determine a speed vector that causes the vehicle to take collision avoidance action based on the passing interval.

[0018] According to the above aspect, collisions can be avoided even when there are multiple passing vehicles.

[0019] FIG. 1 is a block diagram showing an outline of the functional configuration of a vehicle according to a first embodiment. FIG. 2 is a block diagram showing the functional configuration of a collision avoidance device according to the first embodiment. FIG. 3 is a diagram for explaining a collision avoidance function according to the first embodiment. FIG. 4 is a block diagram showing the functional configuration of a collision avoidance device according to a second embodiment. FIG. 5 is a block diagram showing the functional configuration of a collision avoidance device according to a third embodiment. FIG. 6 is a first diagram for explaining a collision avoidance function according to the third embodiment. FIG. 7 is a second diagram for explaining a collision avoidance function according to the third embodiment. FIG. 8 is a block diagram showing the functional configuration of a collision avoidance device according to a modification of the third embodiment. FIG. 9 is a block diagram showing the functional configuration of a collision avoidance device according to a modification of the fourth embodiment. FIG. 10 is a diagram for explaining a collision avoidance function according to the fourth embodiment. FIG. 11 is a schematic block diagram showing the hardware configuration of a collision avoidance device according to at least one embodiment. FIG. 12 is a first diagram showing an example of collision prevention processing in the prior art. FIG. 13 is a second diagram showing an example of collision prevention processing in the prior art.

[0020] First Embodiment Hereinafter, a first embodiment will be described in detail with reference to FIGS.

[0021] (Functional Configuration of Vehicle) Fig. 1 is a block diagram showing an outline of the functional configuration of a vehicle according to the first embodiment. The vehicle V is, for example, an unmanned autonomous mobile body such as an unmanned underwater vehicle, an unmanned guided vehicle, or an unmanned aerial vehicle, or a manned mobile body such as a ship, an aircraft, or a vehicle. An underwater vehicle is also referred to as an underwater vehicle, a ship as a surface vehicle, a carrier vehicle (vehicle) as a ground vehicle, and an aircraft as an aerial vehicle. An unmanned mobile body is also referred to as an unmanned vehicle, and a manned mobile body as a manned vehicle.

[0022] In this embodiment, an example in which the vehicle V is an aerial vehicle will be mainly described. Because an aerial vehicle moves in three-dimensional space, it is suitable for generalizing and explaining vehicle collision avoidance techniques. In other embodiments, the technique described below may be applied to an underwater vehicle, or may be applied to a surface vehicle or a land vehicle by, for example, fixing the up and down movement of the vehicle to the ground or water surface.

[0023] As shown in FIG. 1, a vehicle V includes a collision avoidance device 2, a path following control device 3, and a motion control device 4.

[0024] The collision avoidance device 2 is a device for preventing the vehicle V from colliding with another vehicle (hereinafter also referred to as a "passing vehicle"). Details of the collision avoidance device will be described later.

[0025] The path following control device 3 issues a course command to the motion control device 4 based on a predetermined path of the vehicle V and the position of the vehicle V so that the vehicle V moves along the path.

[0026] The motion control device 4 is a control device including, for example, thrusters and rudders, and adjusts the thrust and rudders of the vehicle V according to the course commanded by the path following control device 3.

[0027] (Regarding normal navigation control of a vehicle) First, the control of normal navigation of the vehicle V will be explained. Note that normal navigation control is known, so only an outline will be explained. The vehicle V has a destination, such as Haneda Airport, and a route to the destination. Even if the navigation of the vehicle V is something like a sightseeing trip to Mt. Fuji, there is a route. The main function of the path following control device 3 is to make the vehicle V follow the route. This function will be called path following control. The path following function is performed by the path control unit 31 setting a route command value p for the vehicle V based on the route and the vehicle position, as shown in FIG. 1. Ｒ The course is the velocity vector of translational motion. The motion control device 4 of the vehicle V adjusts the thrust of the thrusters, the rudder, etc. to control the speed and attitude of the vehicle V, thereby making the course of the vehicle V match the course command value. Under normal circumstances, the vehicle V is controlled in this way.

[0028] (Functional Configuration of Collision Avoidance Device) The collision avoidance device 2 controls the vehicle by determining a separate course command value for collision avoidance and having the motion control device 4 temporarily follow that command value. In other words, when collision avoidance is performed, the original course command value p. Ｒ The original route command value output by the path tracking control device 3 is set to the original route p. ｏｒｇWhen collision avoidance maneuvering begins, the vehicle V deviates from the original path p. ｏｒｇ is basically maintained at the value before the collision avoidance maneuver started. However, for example, there are cases where the vehicle V passes a waypoint on the route during the collision avoidance maneuver. When the waypoint is passed, a course change may be made if the vehicle is normally (during normal navigation). Therefore, even during the collision avoidance maneuver, when an event occurs (such as passing a waypoint) that would normally require a course change, the original course p, ｏｒｇ may be corrected.

[0029] 2 is a block diagram showing the functional configuration of the collision avoidance device according to the first embodiment. As shown in FIG. 2, the collision avoidance device 2 includes an avoidance command candidate generation unit 201, a passing interval estimation unit 202, an original course storage unit 203, and a speed vector determination unit 204.

[0030] The avoidance command candidate generating unit 201 generates an avoidance command candidate group having as elements a plurality of command value candidates for a velocity vector for the host vehicle. For example, the avoidance command candidate group is determined as a table of a plurality of command value candidates, and the avoidance command candidate generating unit 201 generates the avoidance command candidate group by referring to the table recorded in advance in a storage medium.

[0031] The passing interval estimation unit 202 determines, as an avoidance target, a passing vehicle whose relative distance to the host vehicle decreases over time, and estimates the passing interval between the host vehicle and the avoidance target based on the relative speed vector and relative position vector between the host vehicle and the avoidance target. Specifically, the passing interval estimation unit 202 estimates the passing interval between the host vehicle and the avoidance target when the host vehicle performs a collision avoidance maneuver based on the current position vector and speed vector of the host vehicle, a group of relative position vectors between the host vehicle and passing vehicles at the current time, and a group of relative speed vectors between the host vehicle and passing vehicles determined depending on the avoidance command candidate group. Details will be described later.

[0032] The original path memory unit 203 stores the path command value p· of the host vehicle before the collision avoidance was started.Ｒ Alternatively, the velocity vector p· of the own vehicle is stored as the original velocity vector.

[0033] The speed vector determination unit 204 determines, based on the passing interval, a passing vehicle whose relative distance to the host vehicle decreases over time as an avoidance target, and determines a speed vector that allows the host vehicle to avoid a collision with the avoidance target. Specifically, from the group of avoidance command candidates, the speed vector determination unit 204 determines, as the best avoidance command for collision avoidance of the host vehicle, an element that makes the passing interval larger than a specified value and that minimizes the norm of the error between the original speed vector and the speed vector command value candidate. The speed vector determination unit 204 determines a speed vector based on the determined best avoidance command.

[0034] The speed vector determination unit 204 outputs the determined speed vector as a speed command value to the motion control device 4. The motion control device 4 also controls the host vehicle based on the speed command, so that the host vehicle automatically avoids passing vehicles. Note that in another embodiment, the driver may control the motion control device 4 while checking the speed command value output by the speed vector determination unit 204.

[0035] (Collision Avoidance Function) Fig. 3 is a diagram for explaining the collision avoidance function according to the first embodiment. With reference to Fig. 3, the function for avoiding a collision between the own vehicle V and a passing vehicle Vi will be described in detail. When the current time is t=t ０ Figure 3 shows the time t = t ０ In this case, the vehicle V is at position p and speed p, and the passing vehicle Vi is at position q. ｉ , speed q・ ｉ The position and velocity are vector quantities. The origin of the position is O. The position and velocity of each vehicle are expressed in Cartesian coordinates. １ , e ２ , e ３} is the base of the Cartesian coordinate system. ３ is taken vertically upward.

[0036] The origin O of the position may be the current position p of the vehicle V. The position of the passing vehicle Vi is q relative to the vehicle V. ｉ The position p and velocity vector p of the vehicle V can be measured by a sensor mounted on the vehicle V. The position q of the passing vehicle Vi ｉ and velocity vector q ｉ For example, in the case of an aircraft, the relative position and relative speed of the passing vehicle Vi may be obtained from the passing vehicle Vi by a communication means such as Automatic Dependent Surveillance Broadcast (ADS-B). Also, the relative position and relative speed of the passing vehicle Vi may be measured by a radar of the own vehicle V.

[0037] The path T_Vi of the traveling vehicle Vi is determined by the current position q of the traveling vehicle Vi. ｉ and velocity vector q ｉ The path T_V of the vehicle V can be predicted as a straight line determined by the current position p and velocity vector p of the vehicle V. When the velocity vector of the vehicle V is p, the passing interval d (p, p, q) between the vehicle V and a passing vehicle Vi can be predicted as a straight line determined by the velocity vector p of the vehicle V. ｉ , q) is a one-way function as shown in equation (2). d is a non-negative real number. The relative position vector (q ｉ -p) and the relative velocity vector (q ｉ -p·) are parallel, i.e. (q· ｉ -p) × (q ｉ If −p)=0, then d is 0.

[0038]

[0039] In equation (2), the symbol × is the cross product (vector product) operator. The symbol ∘ is used for the inner product (scalar product) operator to distinguish it from the time derivative ∘. If the velocity vector of the subject vehicle V is p·, the path of the subject vehicle V is determined according to p·, and the passing intervals for all passing vehicles Vi from 1 to n (i.e., a group of passing vehicles) can be calculated according to the path of the subject vehicle V. From the viewpoint of collision avoidance, the minimum value among n is important for determining whether the velocity vector p· is good or bad. To be selected, the passing intervals for all passing vehicles Vi from 1 to n must be within the specified interval d. ｉｍｐｏｓｅｄ It has to be bigger.

[0040] Furthermore, the passing related to equation (2) will occur in the future (i.e., t ｘ ≧t ０ ) and those that have already passed each other (i.e., t ｘ <t ０ ) and avoidance commands should be determined from the former. Whether a passing vehicle is about to pass another vehicle or has already passed is checked as follows. If a passing vehicle is about to pass another vehicle, the vehicle and the passing vehicle should be approaching each other, and the sign of the inner product (scalar product) of the relative position vector and relative velocity vector for an approaching object is negative. By using this, it is possible to determine from the group of passing vehicles whether a passing vehicle is about to pass another vehicle (i.e., t ｘ ≧t ０ ) You can select passing vehicles.

[0041] The position of the own vehicle and the positions and speeds of passing vehicles are given for the avoidance maneuver to be performed and cannot be changed. Only the speed vector p of the own vehicle can be changed. In this embodiment, the value of the speed vector is selected from the group of avoidance command candidates {C ｊ p.} (j = 1, 2, ..., m). C ｊ p· is the value of the current velocity vector p· of the own vehicle as the jth avoidance command C ｊ represents the value of the velocity vector changed by

[0042] FIG. B shows an example of avoidance command candidates according to the first embodiment. The avoidance command candidates are determined, for example, by a table such as that shown in FIG. B. The number of avoidance commands is m. In the example of FIG. B, there are 45 candidates, so the value of m is 45. For example, left and right avoidance commands may be generated based on the table in FIG. B. The lane change angle indicated by number 23 is 0 degrees. This indicates that the original lane is maintained in the left and right directions. The candidates may be set so that they are denser the closer to the original lane and sparser the farther from the original lane. This table may also be applied to the up and down directions. If the same table is used for lane changes in the up and down directions, the number of candidates increases to 45 × 45 = 2025. However, since the ascent / descent angle is limited to approximately 20 degrees, the number of candidates may be reduced accordingly. The calculations in this embodiment are one-way functions, and do not require iterative calculations such as those performed by Newton's method. Therefore, even if there are several thousand candidates, the calculation load is not significantly affected.

[0043] In equation (2), p is replaced by C. ｊ Substituting p, the jth avoidance command is C ｊ The passing interval estimation unit 202 evaluates the passing intervals from passing vehicle 1 to passing vehicle n, and calculates the n passing intervals and the predetermined interval d ｉｍｐｏｓｅｄ The minimum difference between the jth avoidance command C ｊ This is expressed mathematically as Equation (3).

[0044]

[0045] Using equations (3) and (4), the velocity vector determination unit 204 determines the velocity vector that will be generated in the future (t ｘ ≧t ０ 2) Passing vehicles are the target of avoidance, and the passing interval and the specified interval d ｉｍｐｏｓｅｄ The difference between the two is positive, and the velocity vector of the original course p ｏｒｇ The best avoidance command is selected as the one with the smallest difference between j and ＊ is an index pointing to the best avoidance command.

[0046]

[0047] Then, the velocity vector determination unit 204 calculates the velocity vector according to the best avoidance command at time t ０ The speed vector command value p ｒｅｆ Let's say.

[0048]

[0049] The collision avoidance device 2 periodically inputs the positions and velocities of the vehicle and passing vehicles, for example, several times per second, and calculates a command value p. of a velocity vector for the vehicle. ｒｅｆ The own vehicle updates the velocity vector command value p. ｒｅｆ Since it follows with a delay of a few seconds, when comparing the instantaneous values, the speed vector p is the command value p of the speed vector. ｒｅｆ However, after a few seconds, p = p ｒｅｆ Based on the current position and speed, the command value p ｒｅｆ If we keep updating, p = p ｒｅｆ At this point, the collision is avoided.

[0050] The velocity vector determination unit 204 may perform the above calculation only when it recognizes the presence of a passing vehicle. If there is no passing vehicle and there is no need to avoid a collision, or if a collision has already been avoided by a collision avoidance maneuver, p = p ｒｅｆ Therefore, collision avoidance maneuvers do not affect the operation of the vehicle.

[0051] (Action, effect) As described above, the collision avoidance device 2 according to this embodiment is equipped with a passing interval estimation unit 202 that targets passing vehicles whose relative distance to the own vehicle decreases over time as a target to be avoided, and that estimates the passing interval between the own vehicle and the target to be avoided based on the relative speed vector and relative position vector between the own vehicle and the target to be avoided, and a speed vector determination unit 204 that determines a speed vector that will allow the own vehicle to avoid collision with the target to be avoided based on the passing interval.

[0052] In this way, even when there are a plurality of passing vehicles, the collision avoidance device 2 can steer the vehicle so as to be able to avoid collision with all passing vehicles.

[0053] Furthermore, the passing interval estimation unit 202 estimates the passing interval with the target to be avoided by performing a cross product operation and an inner product operation expressed as ((relative velocity vector × (relative velocity vector × relative position vector)) ○ relative position vector) ÷ | relative velocity vector × (relative velocity vector × relative position vector) |, where × is the operator of the cross product, 〇 is the operator of the inner product, and | | is the operator of the norm, and determines a passing vehicle to be avoided if the sign of the inner product of the relative position vector and the relative velocity vector is negative.

[0054] In this way, the collision avoidance device 2 can target only passing vehicles approaching the own vehicle as avoidance targets among a plurality of passing vehicles.

[0055] The collision avoidance device 2 further includes an original path memory unit 203 that stores an original speed vector of the host vehicle before collision avoidance began. The passing interval estimation unit 202 estimates a group of passing intervals between the host vehicle and passing vehicles for each element based on the current position vector and speed vector of the host vehicle, a group of relative position vectors with respect to passing vehicles (targets to be avoided) at the current time, and a group of relative speed vectors with respect to passing vehicles determined depending on the group of avoidance command candidates. The speed vector determination unit 204 has an avoidance command determination unit that determines, based on the group of passing intervals, an element that increases the passing interval beyond a specified value and minimizes the norm of the error between the original speed vector and the speed vector command value candidates, as the best avoidance command for the host vehicle, and determines a speed vector for collision avoidance based on the determined best avoidance command.

[0056] In this way, the collision avoidance device 2 can avoid all avoidance targets and minimize deviation from the original course, which means that the effects of collision avoidance, such as an increase in travel distance or travel time, can be reduced.

[0057] Second Embodiment Next, a second embodiment will be described with reference to Fig. 4. Components common to the above-described embodiment will be assigned the same reference numerals, and detailed description thereof will be omitted.

[0058] Aircraft routes are strip-shaped, with vertical separation of 1,000 feet (approximately 300 m) and horizontal separation of 14 nautical miles (approximately 14.4 km). From the perspective of keeping aircraft on route, it is desirable to perform collision avoidance horizontally within the route.

[0059] Therefore, in this embodiment, if a passing vehicle is far away, the risk of collision is not imminent, so route maintenance is prioritized by limiting avoidance commands to a predefined first steering direction (e.g., horizontal direction), and when a passing vehicle approaches, a function is added to prioritize collision avoidance by allowing avoidance commands to a second steering direction (e.g., vertical direction) instead of the first steering direction or in addition to the first steering direction.

[0060] Fig. 4 is a block diagram showing the functional configuration of a collision avoidance device according to the second embodiment. As shown in Fig. 4, the collision avoidance device 2 has a first functional unit 20A that prioritizes course maintenance and a second functional unit 20B that prioritizes collision avoidance. The collision avoidance device 2 further includes a switching unit 205 that switches the command value to be output to the motion control device 4 between the command values ​​for the velocity vector for collision avoidance determined by the first functional unit 20A and the second functional unit 20B.

[0061] The first functional unit 20A has a first avoidance command candidate generation unit 201A, a first passing gap estimation unit 202A, and a first velocity vector determination unit 204A. The first avoidance command candidate generation unit 201A generates a first group of avoidance command candidates that are limited to only avoidance commands for a predetermined first steering direction. In this embodiment, an example will be described in which the first steering direction is the horizontal direction. The functions of the first passing gap estimation unit 202A and the first velocity vector determination unit 204A are the same as the functions of the passing gap estimation unit 202 and the velocity vector determination unit 204 in the first embodiment.

[0062] The second functional unit 20B has a second avoidance command candidate generating unit 201B, a second passing interval estimating unit 202B, and a second velocity vector determining unit 204B.

[0063] The second avoidance command candidate generating unit 201B generates a group of second avoidance command candidates including an avoidance command for a second steering direction different from the first steering direction. In this embodiment, the second avoidance command candidate generating unit 201B generates a group of second avoidance command candidates for changing the command value of the host vehicle in the up / down direction (vertical direction), left / right direction (horizontal direction), and forward / backward direction without imposing any restriction on the steering direction. ２ The second avoidance command candidate (j=1, 2, ..., m ２ ) is referred to as a second group of avoidance command candidates. Note that, in this embodiment, an example will be described in which the second group of avoidance command candidates includes both an avoidance command in the first steering direction (e.g., the horizontal direction) and an avoidance command in the second steering direction (e.g., the vertical direction and the longitudinal direction), but this is not limiting. In other embodiments, the second group of avoidance command candidates may include only an avoidance command in the second steering direction (e.g., the vertical direction).

[0064] The second group of avoidance command candidates is used when the risk of collision with a passing vehicle is imminent (when emergency avoidance is necessary). For example, when the distance to the passing vehicle is d ｉｍｐｏｓｅｄ If the vehicle is close to a passing vehicle, it may be determined that there is an imminent risk of collision (urgent avoidance is necessary). ２ is expressed by equation (6).

[0065]

[0066] The second passing interval estimation unit 202B estimates a group of passing vehicles Q that are at imminent risk of collision. ２ and a second group of avoidance command candidates (j=1, 2, ..., m ２ ) and calculate the passing interval in the same manner as in the first embodiment.

[0067] The second velocity vector determination unit 204B evaluates the equation (7) for the second group of avoidance commands and selects the best avoidance command (second avoidance command). The difference from the first embodiment is that the second avoidance command is used when a passing vehicle has already approached, so the passing interval is set to the safe interval d ｓａｆｅ (d ｓａｆｅ <d ｉｍｐｏｓｅｄ) and, from the viewpoint of speed, the best option is the one that satisfies the safety interval and is closest to the current velocity vector of the vehicle.

[0068] The performance of the second avoidance command is calculated by the passing distance and the safe distance d as shown in equation (7). ｓａｆｅ It is measured by the difference between

[0069]

[0070] The second velocity vector determination unit 204B selects the best avoidance command that has the smallest difference from the current velocity vector p, in other words, the avoidance command that will successfully avoid the collision with the smallest course change, as shown in equation (8). ＊２ is an index pointing to the best second avoidance command.

[0071]

[0072] Then, the second velocity vector determination unit 204B determines the command value p· of the second velocity vector. ｒｅｆ２ is determined by equation (9).

[0073]

[0074] The command value p of the second velocity vector ｒｅｆ２ is a group of vehicles Q that are at imminent risk of collision. ２ Therefore, when there is no vehicle on the road with an imminent risk of collision, i.e., Q ２ If is an empty set, the command value p of the second velocity vector ｒｅｆ２ is not calculated.

[0075] The command value p of the second velocity vector ｒｅｆ２ When the speed vector command value p· is calculated, the switching unit 205 outputs the speed vector command value p· to the motion control device 4. ｒｅｆ , the command value p of the second velocity vector ｒｅｆ２ The command value p of the second velocity vector is selected. ｒｅｆ２ If the velocity vector command value p· is not calculated, the switching unit 205 outputs the velocity vector command value p· to the motion control device 4. ｒｅｆ , the command value p· of the first velocity vector calculated by the first functional unit 20A. ｒｅｆ１ Select the following.

[0076] As described above, the collision avoidance device 2 according to this embodiment adopts, as a command value, a first velocity vector that causes the host vehicle to take evasive action in a predetermined first steering direction when there is no imminent risk of collision of the host vehicle (emergency avoidance is not necessary), and preferentially adopts a second velocity vector that allows evasive action in a second steering direction when there is an imminent risk of collision of the host vehicle (emergency avoidance is necessary). In this manner, the collision avoidance device 2 can, for example, cause the host vehicle to perform collision avoidance maneuvers while maintaining the original course as much as possible when a passing vehicle to be avoided is far away. On the other hand, the collision avoidance device 2 can cause the host vehicle to quickly perform collision avoidance maneuvers when a passing vehicle to be avoided approaches.

[0077] Third Embodiment Next, a third embodiment will be described with reference to Figures 5 to 7. Components common to the above-described embodiments are given the same reference numerals, and detailed description thereof will be omitted.

[0078] Regarding the passage of ships, Japan's Maritime Collision Prevention Act stipulates that "a ship that sees another ship on its right side must keep out of the other ship's path." In this embodiment, for example, a collision avoidance function is added while adhering to this stipulation. Here, "a ship that sees another ship on its right side" refers to a passing vehicle whose line of sight angle is to the right of the own vehicle's velocity vector, and the own vehicle must avoid this passing vehicle.

[0079] 5 is a block diagram showing the functional configuration of a collision avoidance device according to the third embodiment. As shown in FIG. 5, the collision avoidance device 2 according to this embodiment further includes a priority determination unit 206.

[0080] The priority determination unit 206 determines whether each passing vehicle Vi is a passing vehicle that the subject vehicle V should avoid (has priority over the subject vehicle V). Here, as in the above example, a passing vehicle Vi on the right side of the subject vehicle V is determined to be a passing vehicle that the subject vehicle V should avoid. Whether the passing vehicle Vi is on the right side of the subject vehicle V (TRUE or FALSE) is determined by P ｒｉｇｈｔ，ｉ Then, Pｒｉｇｈｔ，ｉ can be evaluated by equation (10). The calculation of the avoidance command is ｒｉｇｈｔ , ｉ In equation (10), sgn is a sign function, and when the input is positive, the output is 1, and when the input is negative, the output is -1. Equation (10) is expressed as P by the product of two sign functions. ｒｉｇｈｔ It stipulates:

[0081]

[0082] Next is "keeping out of the way of other ships." To keep out of the way of other passing vehicles, the own vehicle V should give way to the passing vehicle Vi and pass behind the passing vehicle Vi. To do this, the own vehicle V should change its course so that the point where it will pass the passing vehicle Vi is in front of the own vehicle V.

[0083] The passing interval between the subject vehicle V and a passing vehicle Vi can be measured using equation (2). Equation (2) only measures the length of the passing interval, and cannot distinguish whether the passing occurs in front of or behind the subject vehicle V. Therefore, an evaluation equation for distinguishing between the front and the rear is added. If the passing point between the passing vehicle Vi and the subject vehicle V is in front of the subject vehicle V, the following inequality (11) holds. e ３ is the base of the Cartesian coordinate system, and its direction is vertically upward.

[0084]

[0085] Using this inequality, the signed passing distance d ｓｉｇｎｅｄ is defined by equation (12).

[0086]

[0087] 6 is a first diagram for explaining the collision avoidance function according to the third embodiment. In FIG. 6, a vertically upward base e ３ and the velocity vector p of the vehicle V. The vehicle V passes the passing vehicle Vi on a circle at a distance r, and its position is determined by the base e ３ Let the angle be φ from

[0088] 7 is a second diagram for explaining the collision avoidance function according to the third embodiment. ｓｉｇｎｅｄ This shows the difference between d and d in equation (2). d in equation (2) does not depend on φ. The value of d is constant with r. On the other hand, the signed passing distance d in equation (12) ｓｉｇｎｅｄ has a value of +r if φ is in the open interval (0,π), and a value of -r if φ is in the open interval (-π,0). ｉｍｐｏｓｅｄ While ensuring safety, the vehicle must avoid the path of other ships. ｓｉｇｎｅｄ -d ｉｍｐｏｓｅｄ The velocity vector of the vehicle is selected so that is positive.

[0089] Common sense would dictate, for example, the specified interval d ｉｍｐｏｓｅｄ If you cross the path (ahead) of another vessel while maintaining a distance of several times the distance, you will not be obstructing the path of the other vessel. ｓｉｇｎｅｄ The absolute value of the specified interval d ｉｍｐｏｓｅｄ If the distance between the two vessels exceeds five times the distance between the two vessels, it may be reasonable to allow the vessel to pass in front of the other vessel. Taking this into consideration, the signed passing interval may be evaluated using the following equation (13):

[0090]

[0091] In addition, "prescribed interval d ｉｍｐｏｓｅｄ Five times (5d ｉｍｐｏｓｅｄ ) is an example and may be changed arbitrarily depending on the vehicle's dynamic performance, regulations, etc.

[0092] In this embodiment, the passing interval estimation unit 202 performs a passing interval estimation process that takes priority into consideration. In the passing interval estimation process that takes priority into consideration, a group of passing vehicles that have priority over the own vehicle, i.e., P ｒｉｇｈｔ The value of is TRUE and the signed passing distance d ｓｉｇｎｅｄ Evasion command C ｊ Evaluate the performance of p. ｒｉｇｈｔ The value of is TRUE and the approaching passing vehicles are evaluated by the following equation (14).

[0093]

[0094] Evasion Command C ｊ The score of p is determined by the following formula (15).

[0095]

[0096] The subsequent processing, for example, the processing by the velocity vector determination unit 204, is the same as in the first embodiment.

[0097] As described above, the collision avoidance device 2 according to this embodiment further includes a priority determination unit 206 that determines, for each passing vehicle, whether or not the passing vehicle has priority in the route relative to the own vehicle. The passing interval estimation unit 202 estimates the passing interval for passing vehicles that are determined to have priority, and targets those vehicles to be avoided. In this manner, when priority in the route is determined by law or regulation, the collision avoidance device 2 can avoid a collision between the own vehicle and another vehicle while maneuvering in accordance with the law.

[0098] In this embodiment, the method of determining priority based on the line-of-sight angle has been described using the rule that "a vessel that sees another vessel on its right must keep out of the path of the other vessel" as an example. Note that "a vessel that sees another vessel on its right" is an example of a determination criterion by the priority determination unit 206. For example, even when an overtaking vessel keeps out of the path of the vessel being overtaken, priority is determined based on the line-of-sight angle, and the priority determination unit 206 may use this as a determination criterion.

[0099] <Modification of the Third Embodiment> In the third embodiment, the collision avoidance device 2 determines whether the subject vehicle V is a passing vehicle V i (i=1, 2, . . . , n) that is the subject vehicle V i of the passing vehicles Vi. ｒｉｇｈｔ On the other hand, P ｒｉｇｈｔ If the value of P is FALSE, the vehicle will not be avoided. ｒｉｇｈｔ If the value of is FALSE, the other vehicle will start avoiding the collision first, so the own vehicle does not have to start avoiding the collision. However, if the other vehicle does not start avoiding the collision or if the other vehicle's avoidance is not sufficient, the own vehicle must also start avoiding the collision.

[0100] Fig. 8 is a block diagram showing the functional configuration of a collision avoidance device according to a modification of the third embodiment. In consideration of an emergency, for example, as shown in Fig. 8, the configuration of the priority determination unit 206 of the third embodiment may be added to the configuration of the second embodiment. In this case, if the passing vehicle to be avoided is far away, the determination result of the priority determination unit 206 is input to the first passing interval estimation unit 202A, and it is determined that the passing vehicle has priority over the own vehicle (P ｒｉｇｈｔ ) the value of TRUE.

[0101] On the other hand, as mentioned above, it is not a target for avoidance (P ｒｉｇｈｔ When the risk of collision is imminent as a result of the passing vehicle's insufficient avoidance (the value of which is FLASE), the collision avoidance device 2 may issue the second best avoidance command determined by the second velocity vector determination unit 204B to the subject vehicle V to avoid the nearby vehicle. In this way, even in cases where the subject vehicle has priority and does not normally need to take evasive action, the collision avoidance device 2 can operate the subject vehicle to avoid a collision if the other passing vehicle has not taken sufficient evasive action. In other words, it is possible to more reliably avoid a collision between the subject vehicle and another vehicle.

[0102] Fourth Embodiment Next, a fourth embodiment will be described with reference to Fig. 9. Components common to the above-described embodiments are denoted by the same reference numerals, and detailed description thereof will be omitted.

[0103] It is common for long vehicles such as ships and trucks to pass each other on waterways and roads. For example, a lane on a road is generally shorter than the length of a truck, but as long as the truck stays within the lane, it can pass each other in the opposite direction. Passing each other in the opposite direction is an exceptional method of passing that assumes that the speed vectors of the vehicle and the other vehicle are kept parallel.

[0104] Fig. 9 is a block diagram showing the functional configuration of a collision avoidance device according to a modified example of the fourth embodiment. Fig. 10 is a diagram for explaining the collision avoidance function according to the fourth embodiment. For example, as shown in Fig. 10, if the set value of the passing interval is made variable according to the angle formed by the speed vector of the vehicle and the speed vector of a passing vehicle, two-way traffic becomes possible. Specifically, for example, in the passing interval estimation unit 202 of the first embodiment, d in Equation (4) ｉｍｐｏｓｅｄ As shown in Figure 10, the value of d is changed according to the angle ∠q - ∠p formed by the speed vector p of the own vehicle and the speed vector q of the passing vehicle. Note that this is just one example, and in addition to the angle ∠q - ∠p formed by the speed vector p of the own vehicle and the speed vector q of the passing vehicle, for example, the value of d may be changed based on the magnitude of the relative speed between the own vehicle and the passing vehicle. ｉｍｐｏｓｅｄ The value of may be set.

[0105] In this embodiment, the specified interval d ｉｍｐｏｓｅｄ The value of is not uniform for all passing vehicles, but differs for each passing vehicle. ｉｍｐｏｓｅｄ，ｉ Then, the j-th avoidance command C ｊ The performance of is expressed by the formula (16). The inequality of the formula (16) is ｉｍｐｏｓｅｄ In other words, the inequality in equation (16) expresses that the relative distance between the passing vehicle Vi and the subject vehicle V and the specified interval d ｉｍｐｏｓｅｄ This indicates that the time change of the difference between the specified interval d ｉｍｐｏｓｅｄ is the velocity vector q of the passing vehicle Vi ｉ Since it depends on the angle formed by the velocity vector p of the own vehicle V, the specified interval d ｉｍｐｏｓｅｄ must be treated as time-varying.

[0106]

[0107] Passing interval and specified interval d ｉｍｐｏｓｅｄ The difference between the two is positive, and the velocity vector of the original course p ｏｒｇ The best avoidance command is the one that minimizes the difference between the first and second embodiments.

[0108] As described above, in the collision avoidance device 2 according to this embodiment, the specified distance d ｉｍｐｓｅｄ is the velocity vector p of the vehicle V and the velocity vector q of the passing vehicle Vi. ｉ The collision avoidance control unit 2 controls the vehicle V to pass a passing vehicle in parallel while avoiding a collision with the passing vehicle, without causing the vehicle V to make unnecessary course changes, in a location where two vehicles can pass each other in a two-way traffic manner.

[0109] Other Embodiments Although one embodiment has been described in detail above with reference to the drawings, the specific configuration is not limited to the above, and various design modifications are possible. That is, in other embodiments, the order of the above-described processes may be changed as appropriate. Furthermore, some processes may be executed in parallel.

[0110] <Hardware Configuration> FIG. 11 is a schematic block diagram showing the hardware configuration of the collision avoidance device.

[0111] The computer 800 includes a processor 801 , a main memory device 802 , an auxiliary memory device 803 , and an interface 804 .

[0112] The above-described collision avoidance device 2 is implemented in a computer 800. The operations of the above-described processing units are stored in the form of a program in an auxiliary storage device 803. The processor 801 reads the program from the auxiliary storage device 803, loads it into the main storage device 802, and executes the above-described processing in accordance with the program. The processor 801 also allocates storage areas in the main storage device 802 corresponding to the above-described storage units in accordance with the program.

[0113] The program may be for realizing some of the functions to be performed by the computer 800. For example, the program may be a program that performs functions in combination with other programs already stored in the auxiliary storage device 803 or in combination with other programs implemented in other devices. In other embodiments, the computer 800 may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) in addition to or instead of the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, some or all of the functions realized by the processor 801 may be realized by the integrated circuit.

[0114] Examples of the auxiliary storage device 803 include a hard disk drive (HDD), a solid state drive (SSD), a magnetic disk, a magneto-optical disk, a compact disc read-only memory (CD-ROM), a digital versatile disc read-only memory (DVD-ROM), and a semiconductor memory. The auxiliary storage device 803 may be an internal medium directly connected to the bus of the computer 800, or an external storage device 810 connected to the computer 800 via the interface 804 or a communication line. Furthermore, if the program is distributed to the computer 800 via a communication line, the computer 800 that receives the program may load the program into the main storage device 802 and execute the above-described processing. In at least one embodiment, the auxiliary storage device 803 is a non-transitory tangible storage medium.

[0115] The program may also be a program for realizing part of the above-described functions. Furthermore, the program may be a so-called differential file (differential program) that realizes the above-described functions in combination with another program already stored in the auxiliary storage device 803.

[0116] <Additional Notes> The collision avoidance device, collision avoidance method, and program described in the above-described embodiments can be understood, for example, as follows.

[0117] (1) According to a first aspect, the collision avoidance device 2 is a collision avoidance device 2 that avoids a collision of the subject vehicle with another passing vehicle, and is configured to designate a passing vehicle whose relative distance to the subject vehicle decreases over time as the avoidance target, and is equipped with a passing interval estimation unit 202 that estimates the passing interval between the subject vehicle and the target to be avoided based on the relative speed vector and relative position vector between the subject vehicle and the target to be avoided, and a speed vector determination unit 204 that determines a speed vector that causes the subject vehicle to take collision avoidance action based on the passing interval.

[0118] In this way, even when there are a plurality of passing vehicles, the collision avoidance device 2 can steer the vehicle so as to be able to avoid collision with all passing vehicles.

[0119] (2) According to the second aspect, in the collision avoidance device 2 according to the first aspect, the passing interval estimation unit 202 estimates the passing interval with the target to be avoided by performing a cross product operation and an inner product operation expressed as ((relative velocity vector × (relative velocity vector × relative position vector)) ○ relative position vector) ÷ | relative velocity vector × (relative velocity vector × relative position vector) |, where × is the operator of the cross product, 〇 is the operator of the inner product, and | | is the operator of the norm, and determines a passing vehicle to be avoided as a target to be avoided if the sign of the inner product of the relative position vector and the relative velocity vector is negative.

[0120] In this way, the collision avoidance device 2 can target only passing vehicles approaching the own vehicle as avoidance targets among a plurality of passing vehicles.

[0121] (3) According to the third aspect, the collision avoidance device 2 according to the first or second aspect further includes an original path memory unit 203 that stores the original velocity vector of the host vehicle before collision avoidance began, and an avoidance command candidate generation unit 201 that generates a group of avoidance command candidates whose elements are multiple candidate command values ​​for the velocity vector for the host vehicle. The passing interval estimation unit 202 estimates a group of passing intervals between the host vehicle and the target to be avoided for each element based on the position vector and velocity vector of the host vehicle at the current time, a group of relative position vectors with the target to be avoided at the current time, and a group of relative velocity vectors with the target to be avoided that are determined depending on the group of avoidance command candidates. The velocity vector determination unit 204 determines, based on the group of passing intervals, an element that makes the passing interval larger than a specified interval and that minimizes the norm of the error between the original velocity vector and the candidate command value for the velocity vector, as the best avoidance command for the host vehicle, and determines the velocity vector based on the determined best avoidance command.

[0122] In this way, the collision avoidance device 2 can avoid all avoidance targets and minimize deviation from the original course, which means that the effects of collision avoidance, such as an increase in travel distance or travel time, can be reduced.

[0123] (4) According to a fourth aspect, in the collision avoidance device 2 according to the third aspect, the avoidance command candidate generation unit 201 has a first avoidance command candidate generation unit 201A that generates a first group of avoidance command candidates including only avoidance commands in a first steering direction that is predetermined for the host vehicle, and a second avoidance command candidate generation unit 201B that generates a second group of avoidance command candidates including avoidance commands in a second steering direction that is different from the first steering direction, and the passing interval estimation unit 202 has a first passing interval estimation unit 202A that estimates a first group of passing intervals with an avoidance target based on the first group of avoidance command candidates, and a second avoidance command candidate generation unit 201B that generates a second group of avoidance command candidates including avoidance commands in a second steering direction that is different from the first steering direction. The collision avoidance device 2 further includes a switching unit 205 that, when emergency avoidance is required for the host vehicle, prioritizes the second velocity vector as a command value for the host vehicle.

[0124] In this way, when the passing vehicle to be avoided is far away, the collision avoidance device 2 can make the host vehicle steer to avoid a collision while maintaining the original course as much as possible. On the other hand, when the passing vehicle to be avoided approaches, the collision avoidance device 2 can make the host vehicle steer to avoid a collision quickly.

[0125] (5) According to a fifth aspect, the collision avoidance device 2 according to any one of the first to fourth aspects further includes a priority determination unit 206 that determines, for each passing vehicle, whether the passing vehicle has a priority in the route relative to the own vehicle. The passing interval estimation unit 202 estimates a passing interval for each passing vehicle that is determined to have a priority, as a target to be avoided.

[0126] In this way, when priority of route is determined by laws and regulations, the collision avoidance device 2 can avoid a collision between the vehicle and another vehicle while maneuvering in accordance with the laws and regulations.

[0127] (6) According to the sixth aspect, in the collision avoidance device 2 according to the third aspect, the specified interval changes depending on the angle formed by the velocity vector of the own vehicle and the velocity vector of the passing vehicle, or the relative velocity between the own vehicle and the passing vehicle.

[0128] In this way, in a place where vehicles can pass each other in a facing manner, the collision avoidance device 2 can steer the vehicle so as to pass by in parallel while avoiding collision with a passing vehicle without making unnecessary course changes to the vehicle itself.

[0129] (7) According to a seventh aspect, the collision avoidance method is a collision avoidance method for avoiding a collision between the own vehicle and another passing vehicle, and includes the steps of: determining, as the avoidance target, a passing vehicle whose relative distance to the own vehicle decreases over time; estimating a passing interval between the own vehicle and the avoidance target based on a relative speed vector and a relative position vector between the own vehicle and the avoidance target; and determining, based on the passing interval, a speed vector that causes the own vehicle to take collision avoidance action.

[0130] (8) According to the eighth aspect, the program causes the collision avoidance device 2, which prevents the vehicle from colliding with another passing vehicle, to perform the following steps: estimating the passing interval between the vehicle and the object to be avoided based on the relative speed vector and relative position vector between the vehicle and the object to be avoided, and determining a speed vector that causes the vehicle to take collision avoidance action based on the passing interval.

[0131] According to the above aspect, collisions can be avoided even when there are multiple passing vehicles.

[0132] 2 Collision avoidance device 20A First function unit 20B Second function unit 201 Avoidance command candidate generation unit 201A First avoidance command candidate generation unit 201B Second avoidance command candidate generation unit 202 Spacing estimation unit 202A Spacing estimation unit 202B Spacing estimation unit 203 Original course memory unit 204 Speed ​​vector determination unit 204A First speed vector determination unit 204B Second speed vector determination unit 205 Switching unit 206 Priority determination unit 3 Path following control device 31 Course control unit 4 Motion control device V Vehicle

Claims

1. A collision avoidance device that avoids a collision between the vehicle and another passing vehicle, comprising: a passing interval estimation unit that designates a passing vehicle whose relative distance to the vehicle decreases over time as an avoidance target among the passing vehicles, and that estimates a passing interval between the vehicle and the target to be avoided based on a relative speed vector and a relative position vector between the vehicle and the target to be avoided; and a speed vector determination unit that determines a speed vector that causes the vehicle to take collision avoidance action based on the passing interval.

2. The collision avoidance device of claim 1, wherein the passing interval estimation unit estimates the passing interval with the object to be avoided by performing a cross product operation and an inner product operation represented by ((relative velocity vector x (relative velocity vector x relative position vector)) x relative position vector) ÷ | relative velocity vector x (relative velocity vector x relative position vector) |, where the cross product operator is x, the inner product operator is o, and the norm operator is | |, and the passing vehicle for which the sign of the inner product of the relative position vector and the relative velocity vector is negative is designated as the object to be avoided.

3. A collision avoidance device as described in claim 1, further comprising: an original course memory unit which stores an original speed vector of the host vehicle before collision avoidance begins; and an avoidance command candidate generation unit which generates a group of avoidance command candidates whose elements are multiple command value candidates for the speed vector for the host vehicle, wherein the passing interval estimation unit estimates a group of passing intervals between the host vehicle and the target to be avoided for each of the elements based on the position vector and speed vector of the host vehicle at the current time, a group of relative position vectors with the target to be avoided at the current time, and a group of relative speed vectors with the target to be avoided determined depending on the group of avoidance command candidates, and the speed vector determination unit determines, based on the group of passing intervals, an element that makes the passing interval larger than a specified interval and that has the smallest norm of the error between the original speed vector and the speed vector command value candidate as the best avoidance command for the host vehicle, and determines the speed vector based on the determined best avoidance command.

4. The avoidance command candidate generation unit includes a first avoidance command candidate generation unit that generates a first group of avoidance command candidates including only avoidance commands in a first steering direction predefined for the vehicle, and a second avoidance command candidate generation unit that generates a second group of avoidance command candidates including avoidance commands in a second steering direction different from the first steering direction; the passing interval estimation unit includes a first passing interval estimation unit that estimates a first group of passing intervals with the target to be avoided based on the first group of avoidance command candidates, and a second passing interval estimation unit that estimates a second group of passing intervals with the target to be avoided based on the second group of avoidance command candidates; The collision avoidance device according to claim 3, wherein the speed vector determination unit has a first speed vector determination unit that determines a first best avoidance command based on the first group of passing intervals when emergency avoidance of the host vehicle is not required, and determines a first speed vector based on the determined first best avoidance command, and a second speed vector determination unit that determines a second best avoidance command based on the second group of passing intervals when emergency avoidance of the host vehicle is required, and the collision avoidance device further comprises a switching unit that prioritizes the second speed vector as a command value for the host vehicle when the emergency avoidance is required.

5. A collision avoidance device as described in any one of claims 1 to 4, further comprising a priority determination unit which determines, for each of the passing vehicles, whether or not the passing vehicle has a route priority relative to the vehicle itself, and the passing interval estimation unit estimates the passing interval by treating the passing vehicle determined to have the priority as the avoidance target.

6. A collision avoidance device according to claim 3, wherein the specified interval changes according to an angle between the speed vector of the host vehicle and the speed vector of the passing vehicle, or a relative speed between the host vehicle and the passing vehicle.

7. A collision avoidance method for avoiding a collision between the own vehicle and another passing vehicle, comprising the steps of: determining, as an avoidance target, a passing vehicle whose relative distance to the own vehicle decreases over time among the passing vehicles; estimating a passing interval between the own vehicle and the target to be avoided based on a relative speed vector and a relative position vector between the own vehicle and the target to be avoided; and determining a speed vector that causes the own vehicle to take collision avoidance action based on the passing interval.

8. A program for causing a collision avoidance device that prevents a collision between the vehicle and another passing vehicle to execute the following steps: designating as an avoidance target a passing vehicle whose relative distance to the vehicle decreases over time among the passing vehicles, and estimating a passing interval between the vehicle and the target to be avoided based on a relative speed vector and a relative position vector between the vehicle and the target to be avoided; and determining a speed vector that causes the vehicle to take collision avoidance action based on the passing interval.

Images