TWI712966B - Vehicle prediction system and vehicle prediction method - Google Patents

Abstract

The present disclosure relates to a vehicle prediction method, including the following steps. A processor receives location data of a vehicle during a detection period to generate a first moving route. The processor obtains a first segment route from the back end of the first moving route, and calculates a first prediction route by a curve fitting method. The processor combines the first moving route with the first predicted route into a second moving route. The processor obtains a second segment route from the back end of the second moving route, and calculates a second prediction route by a curve fitting method. When the processor determines that the second predicted route is close to a target, generating a warning signal.

Description

Traffic forecasting system and traffic forecasting method

The present disclosure relates to a vehicle prediction system and a vehicle prediction method, in particular to a technology capable of calculating and predicting the potential movement area of a vehicle.

Advances in science and technology are gradually changing people's lives. Taking the transportation industry as an example, road navigation, driving records, automatic driving, Internet of Vehicles... and other technologies have made driving more convenient for drivers. In addition, "road safety" is the most important aspect of every driver. However, since the driver is likely to be unable to confirm the driving conditions of other vehicles in the distance due to the occlusion of surrounding vehicles during driving, it will be difficult to have enough time to react and avoid when a vehicle approaches at high speed.

One aspect of the present disclosure is a vehicle prediction method, which includes the following steps: a processor receives location data of the vehicle during the detection period to generate a first travel route. The first segmented route is obtained from the rear end of the first travel route forward, and the first predicted route is calculated by the curve fitting method. Will The traveling route is combined with the first predicted route to form the second traveling route. The second segmented route is obtained from the back end of the second travel route forward, and the second predicted route is calculated by the curve fitting method. When judging that the second predicted route is close to the target, a warning signal is generated.

Another aspect of the present disclosure is a vehicle prediction system, which includes a detection device and a processor. The detection device is used for receiving location data of the vehicle. The processor is electrically connected to the detection device to generate the first travel route according to the location data of the vehicle. When the processor does not receive the location data of the vehicle within a predetermined time, the processor is also used to obtain a plurality of travel segments in the first travel route, and generate a plurality of potential routes according to the travel segments, Using these potential routes as boundaries, the predicted travel area is obtained.

Another aspect of the present disclosure is a vehicle prediction method, which includes the following steps: the processor receives the location data of the vehicle during the detection period to generate a first travel route. Obtain a plurality of travel segments in the first travel route. According to these travel segments, a plurality of potential routes are generated by the curve fitting method. Using these potential routes as boundaries, the predicted travel area is obtained.

Accordingly, by dividing the known travel route into multiple segments, it will be possible to more accurately predict the area where the vehicle may move. In addition, even if the detection position of the vehicle is temporarily lost, the processor can still predict the travel route of the vehicle by predicting the route near the second half of the known travel route to reduce the probability of collision.

100‧‧‧Transportation Forecast System

110‧‧‧Processor

120‧‧‧Detection device

121‧‧‧Camera unit

122‧‧‧Network Unit

200‧‧‧Cloud Server

A‧‧‧Transportation

B‧‧‧Transportation

C‧‧‧Transportation

R1‧‧‧First route

Ra~Rd‧‧‧Traveling Segment

Ra1~Rd1‧‧‧Potential Route

Rx‧‧‧Predicted travel area

R1‧‧‧First route

R11‧‧‧First Section Route

R12‧‧‧First prediction route

R2‧‧‧The second course of travel

R21‧‧‧Second Section Route

R22‧‧‧Second Forecast Route

R3‧‧‧The third route

R31‧‧‧Section 3 Route

R32‧‧‧The third prediction route

Rx1‧‧‧Actual route

Rx2‧‧‧Predicted route

S501~S507‧‧‧Step

Fig. 1 is a schematic diagram of a vehicle prediction system according to some embodiments of the present disclosure.

Figure 2 is a schematic diagram of a vehicle prediction system according to some embodiments of the present disclosure.

3A to 3D are schematic diagrams of the operation mode of the vehicle prediction method according to some embodiments of the present disclosure.

Figure 4 is a simulation comparison diagram of the vehicle prediction method.

FIG. 5 is a flowchart of a method for predicting a vehicle according to some embodiments of the present disclosure.

Hereinafter, multiple embodiments of the present invention will be disclosed in the form of drawings. For clear description, many practical details will be described in the following description. However, it should be understood that these practical details should not be used to limit the present invention. That is, in some embodiments of the present invention, these practical details are unnecessary. In addition, in order to simplify the drawings, some conventionally used structures and elements will be shown in a simple schematic manner in the drawings.

In this text, when a component is referred to as “connected” or “coupled”, it can be referred to as “electrically connected” or “electrically coupled”. "Connected" or "coupled" can also be used to mean that two or more components cooperate or interact with each other. In addition, although terms such as “first”, “second”, etc. are used herein to describe different elements, the terms are only used to distinguish elements or operations described in the same technical terms. Unless clearly indicated by the context, the terms do not specifically refer to or imply order or sequence, nor are they used to limit the present invention.

Please refer to FIGS. 1 and 2, which are schematic diagrams of the vehicle prediction system 100 according to some embodiments of the present disclosure. The vehicle prediction system 100 includes a processor 110 and a detection device 120. In some embodiments, the processor 110 and the detection device 120 are arranged in the same vehicle A (eg, a vehicle). The detection device 120 is used to receive location information (such as location coordinates) of other vehicles B and C. The processor 110 is electrically connected to the detection device 120 to receive the location information of the vehicles B and C, and generate travel routes for each vehicle B and C respectively.

In some embodiments, the detection device 120 includes a camera unit 121 (such as a camera lens) for capturing images of the vehicles B and C. The detection device 120 or the processor 110 can calculate the positions of the vehicles B and C based on the captured images (for example, through image recognition technology).

In some other embodiments, the vehicle detection system 100 may use the Internet of Vehicles (IoV) to obtain location information of other vehicles B and C. That is, the vehicles A~C are all connected to the cloud server 200 through the Internet, and the location information is uploaded to the cloud server 200 regularly. In this embodiment, the detection device 120 includes a network unit 122 (such as a network card, a wireless transmission module) for connecting to the cloud server 200, and the detection device 120 uses the network device 122 to connect The cloud server 200 obtains location data of other vehicles B and C.

The processor 110 can record the position data of the vehicles B and C in chronological order, taking the record as the trajectory. In addition, the processor 110 can also segment according to the known travel trajectory, and then calculate for each segment to predict the possible future moving directions of the vehicles B and C. Accordingly, will be able to Prevent vehicle A from colliding with other vehicles B and C. The aforementioned technology is called "segmented route prediction" for short. Specifically, under the technical concept of "segmented route prediction", the present disclosure can also include two application methods: "repetitive route estimation" and "predicted route estimation".

The following first explains "duplicate route estimation". As shown in FIG. 1, the processor 110 receives the location data of the vehicle B through the detection device 120 during the detection period (that is, when the user activates the prediction function). The processor 110 sequentially records multiple pieces of received position data to generate a first travel route R1. The processor 110 is also used to obtain a plurality of travel segments Ra, Rb, Rc, Rd in the first travel route R1. According to the travel segments Ra~Rd, the processor 110 can calculate a plurality of potential routes Ra1~Rd1 using a curve fitting method, and use these potential routes Ra1~Rd1 as boundaries to obtain at least one predicted travel region Rx.

For example, the processor 110 obtains a plurality of travel segments Ra~Rd of different lengths from the rear end of the first travel route R1, and then uses a curve fitting method (curve fitting method) according to each travel segment Ra~Rd. ), calculate the corresponding potential route Ra1~Rd1. As shown in Figure 1, the processor 110 calculates a potential route Ra1 according to the travel segment Ra. Since the direction of each potential route is different, if each potential route is regarded as a fan or two sides of a triangle, it can be regarded as the area that the vehicle B may enter subsequently. In some embodiments, the processor 110 also sets the length of each travel segment Ra~Rd to specifically calculate the area of the predicted travel area Rx. If the processor determines that the predicted travel area will be the same as the travel direction of the vehicle A If there is a conflict, a warning message will be generated, prompting vehicle A to change direction to avoid collision.

The aforementioned "repetitive route estimation" is used as vehicle A When the current position of vehicle B can be confirmed, collision can be avoided by this method. As shown in Figure 1, if the vehicle C is located between the vehicles A and B, so that the vehicle A cannot obtain the location of the vehicle B, the processor 110 will perform "predicted route estimation" at this time. For example, after the processor 110 obtains the first travel route R1 in the aforementioned manner, if it determines that the location data of the vehicle B has not been received within a predetermined period of time, it will perform "predicted route estimation".

The following explains the "predicted route estimation": Please refer to Figures 3A and 3B. The processor 110 obtains the first segment route R11 from the back end of the first travel route R1 forward, and calculates a first predicted route using a curve fitting method. R12. The predicted length of the first predicted route R12 can be adjusted arbitrarily, for example: 30 meters. As shown in FIG. 3A, the processor 110 may calculate a position record of a period of time or a distance from the time point of “losing the position data of the vehicle B” as the “first travel route R1”. In some embodiments, the first length ratio between the first segment route R11 and the first travel route R1 is between one-ninth and one-eleventh. In other embodiments, the ideal ratio is one tenth. That is, the first segment route R11 is 10% of the length of the first travel route R1 near the rear end. For example, when the length of the first travel route R1 is 200 meters, the first segment route R11 may be 20 meters.

As shown in FIG. 3C, after predicting the first predicted route R12, the processor 110 combines the first travel route R1 and the first predicted route R12 into a second travel route R2. This means that although vehicle A has not yet detected the location of vehicle B at this time, vehicle A will assume the predicted first predicted route R12 as the "travel direction of vehicle B". Next, the second segmented route R21 is obtained from the rear end of the second travel route R2 forward, and the second pre-determined route Measure route R22. The second length ratio between the second segment route R21 and the second travel route R2 is greater than the first length ratio between the first segment route R11 and the first travel route R1. In some embodiments, the second length ratio is 1/2, that is, the ratio of the second segment route R21 to the second travel route R2 is “1:2”. After calculating the second predicted route R22, if the processor 110 determines that the second predicted route R22 is close to the vehicle A, a warning signal will be generated.

Please refer to Fig. 3D. Similarly, if the second predicted route R22 is predicted, the processor 110 or the detecting device 120 still has not received the location data of the vehicle B (for example, the vehicle B is still used by the vehicle C). Covered). At this time, the processor 110 will continue to predict the subsequent path of the vehicle B. As shown in FIG. 3D, the processor 110 combines the second travel route R2 and the second predicted route R22 into a third travel route R3. Next, the third segmented route R31 is acquired from the rear end of the third travel route R3 forward, and the third predicted route R32 is calculated by the curve fitting method. The third length ratio between the third segment route R31 and the third travel route R2 is greater than the second length ratio between the second segment route R21 and the second travel route R2. In some embodiments, the third length ratio is 2/3, that is, the ratio of the third segment route R31 to the third travel route R3 is "2:3".

Accordingly, since the vehicle prediction system 100 predicts the subsequent trajectory according to the known driving trajectory of the vehicle B (ie, the first travel trajectory), even if the vehicle A cannot detect the position of the vehicle B, it can continue Predict and judge whether the potential movement path of vehicle B is towards vehicle A to avoid collision.

In some embodiments, when the processor 110 uses the curve fitting method to calculate the predicted route, the processor 110 calculates the vehicle B at different time points. The position data of, substitute into the calculation formula of polynomial approximation, exponential approximation, Fourier approximation, Gaussian approximation, power approximation, rational number approximation, sine curve approximation, interpolation approximation or smooth approximation to obtain the predicted route (such as the first predicted route R2).

For example, the calculation formula of "polynomial fitting method" is " p ( x ) = p ₁ x ⁿ + p ₂ x ^{n -1} +...+ p _n x + p _{n +1} ", where x and y are position data The coordinate value. As shown in Figure 3A, when calculating the first predicted route R12, the processor 110 samples 21 location data from the first segmented route R11 (the number of samples can be adjusted arbitrarily), and substitutes them into the calculation formula of the polynomial fitting method, and sorts them into The following formula:

y = f ( x : a ₀ ,a ₁ ,a ₂ )= a ₀ + a ₁ x + a ₂ x ²

a ₁ x+a ₂ x ²]²

a ₁ x + a ₂ x ² ] ²

After sorting out 21 formulas, using the matrix, the coefficients "a0, a1, a2" can be solved, and the first predicted route M12 can be calculated. Since those skilled in the art can understand the mathematical operation of the curve fitting method, it will not be repeated here.

In addition, the aforementioned "repetitive route estimation" and "predicted route estimation" can be used together. For example, when the processor 100 calculates the predicted route R12, and combines the first travel route R1 and the first predicted route R12 into the second travel route R2, the processor 110 may also obtain a plurality of routes from the second travel route R2. Travel segment, and then use curve fitting method to calculate multiple potential routes according to the travel segment to obtain the predicted travel area Rx (ie, "repetitive route estimation") to determine whether the potential travel area of vehicle B will be covered Traffic worker Drive direction with A. Taking figures 1, 3A, and 3B as examples, when the processor 110 calculates the first predicted route R12, the processor 110 can also calculate the predicted travel area Rx of the vehicle B at the same time to determine whether the predicted travel area Rx will touch Vehicle A.

Please refer to Figure 4, which is a simulation diagram of predictions of vehicles A and B at different locations and routes. From the comparison of the actual route Rx1 and each predicted route Rx2 in the figure, it can be seen that the predicted route Rx2 predicted by the technology of the present disclosure is roughly similar to the actual route Rx1.

Please refer to FIG. 5, where the steps performed by the vehicle forecasting system 100 are described in a flowchart as follows. In step S501, the processor 110 determines whether the detection device 120 obtains the location data of the vehicle B during the detection period. If the detection device 120 obtains the location data of the vehicle B, in step S502, the processor 110 records and updates the first travel route. If the detecting device 120 fails to obtain the location data of the vehicle B, in step S503, as shown in FIG. 1, the processor 110 obtains a plurality of travel segments Ra~Rd in the first travel route R1.

Next, in step S504, the processor 110 generates a plurality of potential routes Ra1~Rd1 using a curve fitting method, and uses the potential routes Ra1~Rd1 as boundaries to obtain the predicted travel area Rx. At this time, if the processor 110 determines that the predicted travel area Rx will touch the vehicle A, a warning message will be generated. The aforementioned steps S503 to S504 are "repetitive route estimation".

In step S505, as shown in FIGS. 3A and 3B, the processor 110 obtains the first segmented route R11 from the rear end of the first travel route R1 forward, and calculates the first predicted route R12 using a curve fitting method. In step S506, The processor 110 combines the first travel route R1 and the first predicted route R12 into a second travel route R2.

If the detection device 120 has not obtained the location data of the vehicle B after calculating the second travel route R2, the processor 110 will continue to predict according to the principles of steps S505 to S506 (ie, "predicted route estimation"). In step S507, the processor 110 obtains the second segmented route R21 from the rear end of the second travel route R2 forward, and calculates the second predicted route R22 by the curve fitting method. In addition, if the processor 110 determines that the second predicted route R22 will intersect the travel route of the target (ie, vehicle A), the processor 110 will also generate a warning signal.

In the foregoing drawings, "vehicles" are taken as examples of vehicles A to C, but the present disclosure can also be applied to route prediction of ships or aircraft.

The various elements, method steps, or technical features in the foregoing embodiments can be combined with each other, and are not limited to the order of description or presentation of figures in the present disclosure.

Although the content of the present invention has been disclosed in the above embodiments, it is not intended to limit the content of the present invention. Anyone who is familiar with the art can make various changes and modifications without departing from the spirit and scope of the content of the present invention. Therefore, the present invention The scope of protection of the content shall be subject to the scope of the attached patent application.

A‧‧‧Transportation

B‧‧‧Transportation

C‧‧‧Transportation

R1‧‧‧First route

Ra~Rd‧‧‧Traveling Segment

Ra1~Rd1‧‧‧Potential Route

Rx‧‧‧Predicted travel area

Claims (20)

A method of vehicle prediction, including: Through a processor, receiving position data of a vehicle during a detection period to generate a first travel route; Obtain a first segmented route forward from the back end of the first travel route, and calculate a first predicted route using a curve fitting method; Combining the first travel route and the first predicted route into a second travel route; Obtain a second segmented route forward from the back end of the second travel route, and calculate a second predicted route by curve fitting; and When it is determined that the second predicted route is close to a target, a warning signal is generated. The vehicle prediction method described in claim 1, further including: After obtaining the first travel route, it is determined whether the location data of the vehicle is received within a predetermined time. The vehicle prediction method according to claim 1, wherein a first length ratio of the first segmented route and the first travel route is smaller than a second length of the second segmented route and the second travel route proportion. The vehicle prediction method according to claim 3, wherein the first length ratio is between one-ninth and one-eleventh. The vehicle prediction method described in claim 1, further including: Obtain a plurality of travel segments in the first travel route; According to these travel segments, a plurality of potential routes are generated by curve fitting; and Using these potential routes as boundaries, at least one predicted travel area is obtained. The vehicle prediction method according to claim 5, wherein the method of obtaining the travel segments further includes: The travel segments are obtained forward from the rear end of the first travel route, and the travel segments have different lengths. The vehicle prediction method according to claim 1, wherein the method of calculating the first predicted route further includes: Take the position data of the vehicle at different points in time and substitute it into a calculation formula of polynomial approximation, exponential approximation, Fourier approximation, Gaussian approximation, power approximation, rational number approximation, sine curve approximation, interpolation approximation or smooth approximation to obtain The first predicted route. A vehicle forecasting system, including: A detection device for receiving location data of a vehicle; and A processor is electrically connected to the detection device to generate a first travel route based on the location data of the vehicle, wherein when the processor does not receive the location data of the vehicle within a predetermined time , The processor also It is used to obtain a plurality of travel segments in the first travel route, and generate a plurality of potential routes according to the travel segments, and use the potential routes as boundaries to obtain at least one predicted travel area. The vehicle prediction system according to claim 8, wherein the processor obtains the travel segments from the rear end of the first travel route forward, and the travel segments have different lengths. The vehicle forecasting system as described in claim 8, further comprising: A vehicle including the processor and the detection device, wherein the processor is further used to obtain a first segmented route from the rear end of the first travel route forward to calculate a first predicted route, and The first travel route and the first predicted route are combined into a second travel route; the processor is also used to obtain a second segmented route from the back end of the second travel route forward to calculate a second predicted route . The vehicle prediction system according to claim 10, wherein a first length ratio of the first segmented route to the first travel route is less than a second length of the second segmented route and the second travel route proportion. The vehicle prediction system according to claim 10, wherein the detection device is used to capture an image of the vehicle, and calculate the location data of the vehicle based on the image of the vehicle. The vehicle prediction system according to claim 10, wherein the detection device is used to connect to a cloud server to obtain the location data of the vehicle from the cloud server. A method of vehicle prediction, including: Through a processor, receiving position data of a vehicle during a detection period to generate a first travel route; Obtain a plurality of travel segments in the first travel route; According to these travel segments, a plurality of potential routes are generated by curve fitting; and Using these potential routes as boundaries, at least one predicted travel area is obtained. The vehicle prediction method according to claim 14, wherein the method of obtaining the travel segments further includes: The travel segments are obtained forward from the rear end of the first travel route, and the travel segments have different lengths. The vehicle prediction method described in claim 14, further including: Obtain a first segmented route forward from the back end of the first travel route, and calculate a first predicted route using a curve fitting method; Combining the first travel route and the first predicted route into a second travel route; Get a second segmented route forward from the back end of the second travel route to The curve fitting method calculates a second predicted route; and When it is determined that the second predicted route is close to a target, a warning signal is generated. The vehicle prediction method described in claim 16, further including: After obtaining the first travel route, it is determined whether the location data of the vehicle is received within a predetermined time. The vehicle prediction method according to claim 16, wherein a first length ratio of the first segmented route to the first travel route is smaller than a second length of the second segmented route and the second travel route proportion. The vehicle prediction method according to claim 18, wherein the first length ratio is between one-ninth and one-eleventh. The vehicle prediction method according to claim 16, wherein the method of calculating the first predicted route further includes: Take the location data of the vehicle at different points in time and substitute it into a calculation formula of polynomial approximation, exponential approximation, Fourier approximation, Gaussian approximation, power approximation, rational number approximation, sine curve approximation, interpolation approximation or smooth approximation to obtain The first predicted route.

Images