JP2019191728A - Speed estimation device - Google Patents

Abstract

To obtain a speed estimation device which can estimate a speed of a moving object appropriately and stably with quick reaction in accordance with situations by estimating the speed more sensitively in the case of the speed being varying and estimating the speed hardly varied from a current estimation speed in the case of the speed being stable.SOLUTION: A speed estimation device 1 of this invention includes: an object detection unit 10 which detects a moving object; a measurement error adjustment unit 20 which adjusts a measurement error of a measurement of the moving object detected by the object detection unit 10; an estimation error adjustment unit 30 which adjusts an estimation error of an estimated speed of the moving object detected by the object detection unit 10; and a speed estimation unit 40 which estimates a speed of the moving object by using the measurement error adjusted by the measurement error adjustment unit 20 and the estimation error adjusted by the estimation error adjustment unit 30.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a speed estimation device that estimates the speed of a moving object.

  Conventionally, there is a technique for observing a moving object and estimating a speed from a time series of the observed positions. At that time, speed estimation is performed using a statistical time series filter such as a Kalman filter. However, due to sudden changes in the state of movement and instability of the observation amount, the follow-up of the estimated speed may be low or the estimation may be incorrect.

  In Patent Document 1, “observation information related to the observation target measured by an external observation device is input, the reference setting value of the observation target is changed to the state of the observation target indicated by the prior probability distribution, and the change is made. A posteriori probability distribution that collates a reference set value with the observation information, calculates an observation likelihood based on the collation error, and creates a posterior probability distribution by weighting the state prior probability distribution according to the observation likelihood. When an interpolation part is detected by performing threshold processing on the weights of a plurality of particles represented by the posterior probability distribution and the estimation part, based on the past state information and the posterior probability distribution, the storage part A technique of “a state estimation device including a defect locus estimation unit that calculates state information” is shown.

JP 2012-3720 A (Patent No. 5436351)

  The technique of Patent Document 1 aims to improve the follow-up of the estimated speed by interpolation of the speed state reproduction probability and the speed from the history, but it is difficult to obtain a good quality history in advance, and the variety of motions The higher the moving object, the worse the accuracy and the more difficult the application because it does not contribute to speed stability.

  The purpose of the present invention is to estimate the speed more sharply when the speed of the moving object is changing, and to estimate that it is less likely to fluctuate from the current estimated speed when the speed is stable. It is another object of the present invention to provide a speed estimation device capable of performing speed estimation with a quick response.

  The speed estimation apparatus of the present invention that solves the above problems includes an object detection unit that detects a moving object, an observation error adjustment unit that adjusts an observation error of an observation amount of the moving object detected by the object detection unit, and the object Using an estimation error adjustment unit that adjusts an estimation error of the estimated speed of the moving object detected by the detection unit, an observation error adjusted by the observation error adjustment unit, and an estimation error adjusted by the estimation error adjustment unit And a speed estimator for estimating the speed of the moving object.

  According to the speed estimation device of the present invention, if it is determined that the moving state is in a state that changes, the speed is likely to change, and if it is determined that the moving state does not change, the speed is unlikely to change. Processing is performed. Therefore, when the speed of the moving object is changing, the speed is estimated more sensitively, and when the speed is stable, it is estimated that the current estimated speed is less likely to fluctuate, and stable and prompt speed estimation can be performed. . Further features related to the present invention will become apparent from the description of the present specification and the accompanying drawings. Further, problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

The functional block diagram of the speed estimation apparatus concerning embodiment of this invention. The flowchart explaining the adjustment method of the observation error by an observation error adjustment part. The flowchart explaining the adjustment method of the estimation error by an estimation error adjustment part. The flowchart explaining a movement state determination process. The flowchart explaining an estimation error change process.

Next, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a functional block diagram of a speed estimation apparatus according to an embodiment of the present invention.
The speed estimation device 1 is mounted on a vehicle such as an automobile, and is realized by cooperation of hardware such as a computer and software. The speed estimation apparatus 1 observes moving objects such as pedestrians and other vehicles around the host vehicle. Then, the observation error corresponding to the weight at the time of estimation for the observation amount of the moving object and the estimation error corresponding to the weight at the time of estimation of the estimated speed of the moving object are adjusted. Then, the velocity of the moving object is estimated using the adjusted observation error and estimation error. By adjusting the observation error and the estimation error according to the state of the moving object, the speed is estimated more sharply when the speed of the moving object is changing, and when the speed is stable, it varies from the current estimated speed. It is difficult to estimate, and it is possible to perform speed estimation appropriately and stably in response to the scene.

  The speed estimation apparatus 1 includes an object detection unit 10, an observation error adjustment unit 20, an estimation error adjustment unit 30, and a speed estimation unit 40. Further, as a storage unit, an observation history table 50 and an estimation history table 60 are stored. And an observation error table 70. The observation history table 50 stores the history information of the observation amount of the moving object, the estimation history table 60 stores the history information of the estimated speed of the moving object and its error, and the observation error table 70 stores the moving object. The information of the observation error of the observed quantity is stored.

  In the speed estimation device 1, the object detection unit 10 detects a moving object around the host vehicle. At the same time, the position of the moving object is specified. The observation error adjustment unit 20 refers to the observation history table 50, the estimation history table 60, and the observation error table 70 to adjust the observation error of the observation amount of the moving object. Then, the estimation error adjustment unit 30 refers to the estimation history table 60 and the observation error table 70 to adjust the estimation error of the estimated speed of the moving object. Then, the speed estimation unit 40 estimates the speed of the moving object using the observation error and the estimation error.

<Object detection unit 10>
The object detection unit 10 detects a moving object such as a pedestrian or other vehicle around the host vehicle using a known technique such as a camera or a radar. Then, by detecting the moving object every predetermined time, the actual position and moving speed of the moving object can be grasped. The observation amount that is the detection result of the moving object is stored in the observation history table 50 as an observation history.

<Observation error adjustment unit 20>
FIG. 2 is a flowchart for explaining an observation error adjustment method by the observation error adjustment unit.
The observation error adjustment unit 20 sets the observation error of the moving object according to whether or not the moving object is a target for which the observation error has been measured in advance, and observes the history estimation from the observation position history and the estimated position history of the moving object. Calculate the error. Then, the set observation error is corrected using the calculated observation error of the history estimation.

  First, in order to set the observation error of the moving object, the brightness environment of the image is determined (S201). The determination of the brightness environment of the image is performed using the exposure time, gain, and luminance value of the image. Then, it is determined whether or not the moving object has been identified and is an identification target whose observation error has been measured in advance (S203). Whether or not it is an identification target is determined by referring to the observation error table 70.

  The observation amount of the moving object is three-dimensional data. For example, when the moving object is a pedestrian, the rectangular area (i, j) surrounding the pedestrian on the image and the parallax average value (from the own vehicle) Distance to pedestrians). However, representative points such as the center of gravity of the rectangular area are used for speed estimation. That is, the image coordinates of the representative points and the parallax average value can be used as the observation amount. Naturally, these observation amounts include errors (observation errors), and the amount of the error also differs between pedestrians and vehicles, trees, grass, and the like.

  In the observation error table 70, it is useful to identify a moving object as much as possible and measure and store a corresponding observation error in advance. By storing a plurality of observation errors, if the target is suitable in the determination process of S203, the observation error of the suitable target can be set.

  For example, the observation error is measured by locating a pedestrian at a distance of 20 m from the camera and a representative point at a position 600 pixels from the left on the image. This can be done by calculating the variance of the observed amount over a certain period of time and estimating the observation error. The observation error can be provided for each position of the object or may be changed depending on the background of the object. Representative values such as the average of many assumed scenes can also be used.

  In the determination process of S203, when it is determined that the moving object is the identification target that has been measured (YES in S203), the appearance angle, posture shape, and color of the moving object that are assumed in advance are determined (S204). . There may be one kind of posture shape. For example, in the case of a pedestrian, as an example of the viewing angle, side orientation, frontal orientation, back orientation, etc. are determined, and examples of posture shapes include arm swing, both hands, waving one hand, upper body is For example, the color of the clothes is determined as an example of the color. Therefore, for example, it is determined that the pedestrian is (1) waving his hand, (2) his back, (3) his upper body is looking forward, and (4) his clothes are black.

  Then, an observation error that has been measured in advance according to the brightness of the image, the identification target, the viewing angle, the posture shape, and the color is set (S205). These observation errors are stored in the observation error table 70 and are set by referring to the observation error table 70. For example, if the identification target is a pedestrian, the viewing angle is sideways, and the color is black, the observation error is large because the edge is relatively small on the image and the edges are difficult to detect. Is set to

  On the other hand, if it is determined in S203 that the measurement target is not already identified, an observation error that has been measured in advance according to the brightness environment of the image is set (S206). Here, since it is determined that the object is not already measured, the only clue for setting the observation error is the brightness environment. Therefore, the observation error is set by referring to the observation error table 70 using the brightness environment determined in S201. For example, when the surroundings are dark at night as a bright environment, it is difficult to specify an area on the image of the target object, so that the observation error is set to be larger than that in a bright daytime.

  On the other hand, in step S202, the observation error of the history estimation is calculated from the observation position history and the estimated position history of the moving object. The observation position history and the estimated position history are stored in the observation history table and the estimation history table, respectively. The observation position history is history information of the position of the actually detected moving object, and the estimated position history is history information of the position of the estimated moving object. Here, based on the estimation result, when the moving speed of the moving object is a constant speed, the deviation between the estimated position and the observation position is interpreted as an observation variation, and the standard deviation of the deviation set for a certain time is estimated as an observation error. As a condition for calculating the observation error of history estimation, calculation is not performed unless the speed of the moving object is constant (including speed 0). That is, when the speed of the moving object is changing, the observation error of the history estimation is not calculated. Whether the speed is constant or not is, for example, constant speed when the most recent estimation result for a certain time is almost constant (when the speed changes within the threshold from the average of the estimated values for a certain time). It is judged that.

  Then, correction based on observation errors in history estimation is performed (S207). Here, processing for correcting the observation error set in S205 or S206 using the observation error of history estimation calculated in S202 is performed. When the moving speed of the moving object is not constant and the observation error for history estimation is not calculated in S202, the correction process for the observation error set in S205 or S206 is not performed. In such a case, the observation error is not arbitrarily manipulated. If the observation error set in S205 or S206 is, for example, more than twice or half of the observation error in the history estimation in S202, for example, the observation error set in S205 or S206 and the history estimation in S202 Correction that averages the observation error may be performed. The observation error adjusted by the correction is stored in the observation error table 70. The observation error adjustment unit 20 does not perform the process of changing the observation error unless there is special information such as a constant speed.

  FIG. 3 is a flowchart illustrating a method for adjusting an estimation error by the estimation error adjustment unit, FIG. 4 is a flowchart illustrating a movement state determination process, and FIG. 5 is a flowchart illustrating an estimation error change process.

  The estimation error adjustment unit 30 includes a movement state determination unit and an estimation error change unit, and performs a movement state determination process (S301) and an estimation error change process (S302) as shown in FIG. In the movement state determination process (S301), a process of determining whether the movement state of the moving object is a preset acceleration / deceleration state, a constant speed state, or a general state is performed.

  Here, the acceleration / deceleration state is defined as a state where acceleration is unlikely to be zero or a state where acceleration / deceleration is likely. Therefore, the acceleration / deceleration state includes not only a state where acceleration / deceleration is actually performed, but also a case where acceleration / deceleration is predicted. For example, even when the vehicle is currently stopped (even when the speed is 0), if the movement of the moving object is predicted due to the change from the red signal to the blue signal, the acceleration / deceleration state is determined. By grasping the acceleration / deceleration state, an accurate speed is promptly estimated from the position information where the position change has occurred due to actual acceleration / deceleration.

  The constant speed state is defined as a state where acceleration is zero. The constant speed state includes a case where the speed is zero. Therefore, the state predicted to be stopped is also included. By grasping this, a meaningless speed is not estimated due to a slight positional deviation.

  The general state is defined as a state (an unknown state) in which neither the acceleration / deceleration state nor the constant speed state is determined. In such a case, the estimation error is not arbitrarily manipulated.

  As shown in FIG. 4, the moving state of the moving object is <1. Works>, <2. Estimated speed history>, <3. It is estimated by at least one of the specific positions>, and is determined by voting these three kinds of estimation results. Then, the determination result and the reason for the state determination are used in the next estimation error changing process (S302). The priority order is set in advance for the information on the action of the moving object, the information on the estimated speed history of the moving object, and the information on the specific position of the moving object. It is estimated (S402, S403, S406).

<1. Works>
When the action of the moving object is recognized (S401), the movement state is estimated based on the action (S402). The action of the moving object is recognized using a known technique, for example, analyzing an image captured by a camera. The estimation of the movement state in S402 is estimated by an operation having a high priority. For example, in the case of a pedestrian, the priority order is (1-1) sitting at a constant speed, (1-2) accelerating / decelerating when waving his face, and (1-3) self-orientation. It is set in the order of the acceleration / deceleration state that is stopped in the direction crossing the front of the vehicle.

  In (1-1) above, you cannot move because you are sitting. In (1-2), you are trying to cross and start moving. In (1-3), you may be trying to move across. It is predicted. Therefore, in the state of (1-1), it is expected that the state of (1-2) and (1-3) does not move even if the state of (1-2) and (1-3) is established, so the priority order becomes the highest. In the state (1-2), it is predicted that the vehicle will cross even if (1-3) does not hold, so the priority order is second.

  Further, for example, in the case of a vehicle, (2-1) the acceleration / deceleration state during running by turning on the hazard lamp, (2-2) the constant speed state during stopping by turning on the hazard lamp, (2-3) direction It is determined which state is the acceleration / deceleration state in which the indicator is lit.

  In the above (2-1), since the hazard lamp is turned on, it will be stopped. In the above (2-2), the hazard lamp is turned on, so it will be stopped as it is. ), It is predicted that the direction indicator will be lit and will move in that direction. Since the states (2-1) to (2-3) are exclusive, there is no priority order and they are set out of order.

<2. Estimated speed history>
The moving state is estimated from the estimated speed history (S403). The estimated speed history is read from the estimated history table 60. The estimated speed history includes information on the estimated speed and an error of the estimated speed. The estimation of the movement state in S403 is performed from a history with high priority. For example, (3-1) an acceleration / deceleration state in which the amount of change in the estimated speed is sufficiently larger than the estimated speed error (for example, the amount of change is 1σ or more of the estimated speed), and (3-2) the estimated speed history is They are set in the order of a constant speed state that is sufficiently constant for a sufficiently long time (continuous).

  In (3-1) above, it can be determined that the speed has changed to the same extent as the estimation error and is accelerating rapidly. In (3-2) above, since it is actually constant, it will continue to be constant. , Is predicted. Therefore, even if the above (3-2) is established except for the most recent time, the above (3-1) may occur suddenly, and at this time, the movement starts, so the above (3-1) is more suitable than the above (3). The priority is higher than -2).

<3. Specific position>
Traffic environment information is acquired (S404). The traffic environment information can be acquired by map information such as car navigation or external recognition by a camera or the like. Then, the specific position of the moving object is determined (S405), and the moving state of the moving object is estimated from the specific position (S406).

  The estimation of the specific position in S406 is determined by the specific position with the highest priority. For example, in the case of a pedestrian, (4-1) an acceleration / deceleration state that collides with another object (still or moving) in the remaining few seconds, and (4-2) a direction in which a guardrail or a wall restricts movement. Constant speed state (speed in that direction is 0), (4-3) acceleration / deceleration state standing in the crossing direction before the pedestrian crossing when the crossing signal is blue (if there is no pedestrian crossing without a signal) The condition of green light is unnecessary). In the state of (4-1), the speed should be changed so as not to collide. In the state of (4-2), the movement in the direction in which the movement is restricted is impossible, so the speed in that direction is 0. This is because in the state (4-3), it is predicted that the vehicle should be traversed by a green light. The state of (4-2) above is an independent term, and it can be handled independently from the others to reduce the estimation error in that direction. Since the states (4-1) and (4-3) are exclusive, they are set in any order.

  For example, in the case of a vehicle, (5-1) an acceleration / deceleration state in which the vehicle is approaching another object (stationary or moving) and collides in the remaining few seconds, (5-2) a guardrail or a wall restricts movement. There are cases where it is a constant speed state in a certain direction, and (5-3) an acceleration / deceleration state where the vehicle stops at the right turn start position in the intersection. In (5-1) above, the speed should be changed so as not to collide, and in (5-2) above, the movement in the direction in which movement is restricted cannot be performed, so the speed in that direction should be 0, This is because in 5-3) it is predicted that it should start suddenly because it is going to turn right. The priority order is the same as the above states (4-1) to (4-3).

  In steps S402, S403, and S406 described above, when the movement state is estimated based on (1) the work, (2) the estimated speed history, and (3) the specific position, a process of determining the movement state by voting is performed (S407). ). In S407, for example, the acceleration / deceleration state is +1, the constant speed state is -1, the general state is 0, and it is determined that the acceleration / deceleration state is +1 or more, the constant speed state is -1 or less, and the general state is 0.

  Next, in the estimation error changing process (S302) shown in FIG. 5, adjustment for changing the estimation error is performed according to the movement state and the determination reason. First, the determination result by the movement state determination of S301 (whether the movement state is (1) acceleration / deceleration state, (2) constant speed state, or (3) general state) is input (S501).

  (1) In the acceleration / deceleration state, processing for making the estimation error of the moving speed larger than the value of the previous speed estimation result is performed (S502). Here, the estimated speed variance is increased by decreasing the estimated speed weight. The value that increases the estimation error of the moving speed is calculated according to the reason determined to be the acceleration / deceleration state, and an average value is used when there are a plurality of reasons. For example, when the speed change is large (estimated from the history), the average value of the speed change amount and the estimated standard deviation is set as a new estimated error (standard deviation). It may be simply a fixed magnification (for example, twice). Further, the value may be adjusted according to the distance from the host vehicle, the position in the lateral direction with respect to the traveling direction of the host vehicle, the identification target, and the current estimation error.

  (2) In the case of a constant speed state, a process of making the estimation error of the moving speed smaller than the value of the previous speed estimation result is performed (S503). Here, the estimated speed weight is increased to reduce the estimated speed variance. The value for reducing the estimation error of the moving speed is calculated according to the reason determined to be the constant speed state, and an average value is used when there are a plurality of reasons. For example, when the speed history is sufficiently stable (estimation from the history), the standard deviation in the time direction calculated from the estimated speed history is set as a new estimation error (standard deviation). It may be simply a fixed magnification (for example, 0.5 times). Further, the value may be adjusted according to the distance from the host vehicle, the position in the lateral direction with respect to the traveling direction of the host vehicle, the identification target, and the current estimation error.

  (3) In the case of the general state, no adjustment for changing the estimation error is performed. Since the general state is neither the acceleration / deceleration state nor the constant speed state, and neither is determined (unknown state), the estimation error is not arbitrarily manipulated in such a case.

  As described above, in the estimation error changing process S302, the value of the estimation error is changed depending on whether the moving state of the moving object is the acceleration / deceleration state or the constant speed state, and the maximum value and the minimum value are set in advance. In addition, the value obtained by changing the estimation error is prevented from becoming excessively small or large, resulting in overestimation.

  Further, in the estimation error changing process S302, if the distance or the lateral position is far and there is no immediate danger, the adjustment for changing the estimation error may be omitted in order to avoid intentional change of the estimation sensitivity. Further, the adjustment may not be performed if the sensitivity that can be sufficiently obtained with the current estimation error can be handled. And you may make it simply larger or smaller by fixed magnification, without dividing according to a movement state and its determination reason. Emphasizing stability estimation, the rate of change, which is the rate at which the estimation error is changed, can be set small. In the case of a constant speed state in one direction due to a narrow road width, etc., only the estimated speed error in that direction is adapted. That is, if the road width is narrow, the vehicle cannot move in the road width direction, so that the error in the estimated speed in the road width direction is reduced.

  According to the estimation error changing process (S302) described above, for example, it is determined that the pedestrian is in a state of turning his / her body on the roadway at a position near the roadway on the sidewalk and looking at the surroundings of the host vehicle with the action of the face. In this case, the acceleration / deceleration state with a high possibility of crossing the roadway and a high possibility of sudden movement is determined. Therefore, it becomes ± 0 in voting, and it is determined as a general state in S407. Therefore, the estimation error is not changed. However, if it is a pedestrian crossing with a green light, it is determined to be in an acceleration / deceleration state because of a specific position. Accordingly, a change process for increasing the estimation error is performed in S502. Further, if the vehicle is moving moderately, it is not determined that the vehicle is in a constant speed state from the speed history, and similarly, a change process for increasing the estimation error is performed.

  The degree to which the speed estimation error is adjusted is changed in accordance with the desired reaction speed, processing cycle, and target speed. For example, if the processing cycle is 0.01 seconds and you want to calculate the correct speed within 0.2 seconds with respect to the pedestrian's launch, it is possible to estimate 20 times, so it will follow the pedestrian's speed change sufficiently. Therefore, the change in error is very small, for example, a change of 5% at a time.

  According to the speed estimation device 1 of the present embodiment described above, by estimating the speed by adjusting the observation error and the estimation error of the moving object, the speed is estimated more sharply when the speed of the moving object is changing. When the speed is stable, the current estimated speed is not easily changed. Therefore, it is possible to perform speed estimation appropriately and stably according to the scene.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to the above-described embodiments, and various designs can be made without departing from the spirit of the present invention described in the claims. It can be changed. For example, the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to one having all the configurations described. Further, a part of the configuration of an embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of an embodiment. Furthermore, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

1 speed estimation device 10 object detection unit 20 observation error adjustment unit 30 estimation error adjustment unit 40 speed estimation unit 50 observation history table 60 estimation history table 70 observation error table

Claims (6)

An object detection unit for detecting a moving object;
An observation error adjustment unit that adjusts an observation error of the observation amount of the moving object detected by the object detection unit;
An estimation error adjustment unit that adjusts an estimation error of an estimated speed of the moving object detected by the object detection unit;
A speed estimation unit that estimates the speed of the moving object using the observation error adjusted by the observation error adjustment unit and the estimation error adjusted by the estimation error adjustment unit;
A speed estimation apparatus comprising: The estimated error adjustment unit includes a moving state determination unit that determines whether the moving state of the moving object is a preset acceleration / deceleration state, a constant speed state, or a general state;
An estimation error changing means for increasing an estimation error of the moving object when the determined moving state is the acceleration / deceleration state, and reducing an estimation error of the moving object when the determined moving state is the constant speed state; The speed estimation apparatus according to claim 1, further comprising:   The moving state determination means determines the moving state of the moving object based on at least one piece of information on the action of the moving object, an estimated speed history of the moving object, and a specific position of the moving object. The speed estimation device according to claim 2. Priorities are set in advance for the information on the action of the moving object, the information on the estimated speed history of the moving object, and the information on the specific position of the moving object, respectively.
The speed estimation apparatus according to claim 3, wherein the moving state determination unit determines a moving state of the moving object from information having the high priority.   The observation error adjustment unit sets an observation error of the moving object according to whether or not the moving object is a target for which an observation error has been measured in advance, and estimates a history from the observed position history and the estimated position history of the moving object 5. The velocity estimation according to claim 1, wherein the observation error is calculated, and the set observation error is corrected using the calculated observation error of the history estimation. apparatus.   The speed estimation apparatus according to claim 5, wherein the observation error adjustment unit calculates an observation error of the history estimation on condition that a moving speed of the moving object is a constant speed.

Images