CN104134378A - Urban rail train intelligent control method based on driving experience and online study - Google Patents
Urban rail train intelligent control method based on driving experience and online study Download PDFInfo
- Publication number
- CN104134378A CN104134378A CN201410283677.0A CN201410283677A CN104134378A CN 104134378 A CN104134378 A CN 104134378A CN 201410283677 A CN201410283677 A CN 201410283677A CN 104134378 A CN104134378 A CN 104134378A
- Authority
- CN
- China
- Prior art keywords
- train
- time
- speed
- speed limit
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention discloses an urban-railway-system train intelligent control method based on driving experience and online study. The method obtains output of a controller through simulation of a driving strategy of an experienced driver, use of real-time data of train operation and reasoning and induction, and uses real-time information and speed-limit information of the trains to reasonably distribute the remaining time of train operation so that a time precision estimation value and a speed estimation value are obtained, and based on an MTDA algorithm of the real-time information, the method uses a train operation limit speed and current position, speed and time to deduct an expectation operation time and an expectation operation speed of train operation. Compared with a simulation result of automatic train operation (ATO) of a Yizhuang line of the Beijing city, the intelligent control method is capable of saving energy and improving on-time performance.
Description
Technical field
The present invention relates to track traffic control field, relate more specifically to a kind of urban track traffic (city rail) train intelligent control method based on driving experience and on-line study.
Background technology
Train automatic controlling system ATC (Automatic Train Control) adopts advanced train automatic control technology can greatly improve efficiency, the security of driving, for the requirement of urban track traffic high-efficiency high-density, is absolutely necessary.ATC system generally comprises ATS (Automatic Train Supervision) system, ATP (Automatic Train Protection) system and ATO (Automatic Train Operation).Wherein, ATO, as one of train automatic controlling system important subsystem, utilizes vehicle-mounted curing information and terrestrial information to realize the control to train traction, braking, make train often in optimal operational condition, improve passenger's comfort level, improve train punctuality rate, save the energy.Concerning urban track traffic operation control system, no matter be pilot steering or omnidistance unmanned, ATO will bring into play its vital role, and the train running speed between two stations is controlled, and it controls effect directly affects property indices.In ATO, apply different control algolithms, it controls effect is different, therefore, is necessary to study effective ATO control algolithm, so that train is to greatest extent in optimal operational condition.
In practical application, ATO control algolithm is that PID controls the most widely.PID is a kind of linear regulator, it by the deviation of setting value and output valve in proportion, integration and differentiation controls.This control method will be set out distance-rate curve in advance.Obviously, this is a kind of method of carrying out speed control according to the driving curve of prearranging.Acceleration-deceleration switching times when the shortcoming of this method is control rate is too much, this situation had both been unfavorable for even running, destroy again the comfortableness of taking, also increased energy consumption simultaneously and reduced stopping accuracy, and cannot adjust flexibly according to operation planned time; The frequent switching of this outer controller has also affected the serviceable life of controller.
In addition ATO control algolithm also comprises the methods such as adaptive robust control, iterative learning control, fuzzy control, and still, these methods all need the target velocity curve of off-line, is difficult to take into account the problems such as the comfort of passenger in train travelling process; And these control algolithms are difficult to the train model of thorough consideration complexity, and in train travelling process, can be subject to needing the interference of X factor, as the problem such as nonlinear characteristic and saturation characteristic of the impact of the factors such as the acting force between uncertain friction force, compartment, the output of Train Control device, very difficult realization is accurately followed the trail of, and is difficult to realize the driving strategy of online-intelligence.
Summary of the invention
For avoiding above the deficiencies in the prior art, the invention provides a kind of urban railway transit train intelligent control method based on driving experience and on-line study.
The present invention adopts following technical proposals:
A urban railway transit train intelligent control method based on driving experience and on-line study, utilizes the real-time information of train operation, realizes the operation of train multiple goal and controls, and this control method comprises the steps:
1), when train operation, every the time interval of Δ t, determine train current location s and present speed v information, and the speed limit (S of circuit
1_b, S
2_blS; V
l_1, V
l_2l0), expect the interval S of running time T and Train Stopping
pif, judgement s>=S
p, turn to step 5, enter shutdown phase, otherwise, turn to step 2;
2) judge the current residing speed restrictive block of train, i.e. S
k-1≤ s≤S
k, k represents the current residing speed restrictive block of train, corresponding speed limit size is V
k;
3) obtain the current operating mode of train according to expert knowledge library, again according to speed limit situation and train present speed, calculate the shortest working time of each speed limit section and remaining runtime is pro rata distributed in each speed limit section, and then drawing the remaining runtime in current speed limit section; Again by inference machine to the current operation information of train infer after this whether train accelerate, deceleration or coasting;
4) train intelligent driving controller is carried out to on-line study, obtain this controller output valve;
5) train operation circuit is arranged to transponder, and the positional information providing according to transponder is adjusted described controller output valve.
Further, in described step 3, the current operation information of train comprises present speed, current location, current speed limit and current residual working time.
Further, in described step 3, it is theorized that machine to the current operation information of train infer after this whether train accelerate, deceleration or coasting specifically comprise:
The speed limit not declining when train front and when stopping, the operating mode of train can only be to accelerate or coasting, now again computing time precision estimation value
wherein T
afor time precision estimated value, T
efor time estimation value, T
cfor the remaining runtime in current speed limit section, just can infer acceleration or coasting operating mode by comparing time precision value and velocity amplitude: if T
a>T
a° and
accelerate wherein T
a° for set maximum time trueness error, the size of acceleration is a=a
c+ Δ a
max, the controller that wherein a is next step is initially exported, a
cfor the output of back controller, Δ a
maxfor the maximal value Δ a of acceleration change
max=0.2; Otherwise coasting, coasting Time Controller is finally output as a=0;
In the time that there is the speed limit of decline in train front, according to ATP curve generation method, in certain distance threshold doseag, slow down to prevent from exceeding speed limit, apart from threshold doseag value be
wherein V
l_nextfor next section speed limit, γ is the number percent that velocity coefficient expects to decelerate to next speed limit, to prevent from exceeding speed limit, and γ=0.9, v is current speed, a
setfor the reference deceleration degree arranging, the size of value should meet comfort level requirement, can not excessive a
set=-0.5m/s
2, within the scope of train operation enters into apart from threshold doseag after, according to the retarded velocity of setting, to slow down be that controller is initially output as a=a
set=-0.5m/s
2.
Further, described step 4 is carried out on-line study to train intelligent driving controller and is comprised:
1) the objective optimization function of setting gradient descent method is f=(V
e-v)
2, be input as the current operation information of train;
2) judge the current residing speed restrictive block of train, i.e. S
k-1≤ s≤S
k, k represents the current residing speed restrictive block of train, corresponding speed limit size is V
k;
3) according to following process computation velocity estimation value
and V
e:
T
L＝T-t；
4) train operation distance and train present speed can be with representing as follows:
s＝∫vdT＝V
avgt；
v＝∫adT＝a
avgt；
Wherein, V
avgfor the average velocity moving before this, a
avgfor the average acceleration of moving before this;
5) bring above formula into objective optimization function, and to the t gradient that obtains objective function of differentiating be:
6), based on gradient descent method, known Train Control device output is adjusted into
λ is step-length, and numerical value is λ=0.01,
for Grad.
Further, when the positional information providing according to transponder in described step 5 is adjusted described controller output valve, the retarded velocity estimated value of estimating during through i transponder is
wherein V
cfor current speed, S
ibe the distance of i transponder apart from parking terminal, by a
e_irunning to the initial output of i+1 the controller between transponder as train from i transponder is that this Time Controller is initially output as a=a
e_i, the like, to the last stop.
Beneficial effect of the present invention is as follows:
The present invention, simulates driver's train intelligent driving technology and can drive online urban railway transit train, does not need the off-line target velocity curve in control method in the past, and can overcome the uncertainty of train model parameter; By summing up Expert Rules, set up expert knowledge library, simulate experienced driver's driving strategy, consider operation and the docking process of train; Utilize the real time data in train travelling process, we propose and have designed MTDA algorithm, to calculate the parameters such as real-time distribution excess time of train; Definition and the on-line study formula of having derived based on gradient method, with the performance index of on-line optimization train travelling process; And we are according to the real data of Beijing Metro Yi Zhuang line, the present invention has been carried out to online emulation experiment, the driving curve that has contrasted the inventive method is driven curve with the PID in actual operation, result shows, the present invention can ensure that train does not exceed the speed limit, and arrives at a station on schedule, and has improved passenger's ride quality, reduce operation energy consumption, met stopping accuracy; In addition we have also carried out emulation for the Train Control method under different expectation conditions working time, and result shows, driving strategy of the present invention can adapt to different expectation working times.
Brief description of the drawings
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail.
Fig. 1 illustrates structure principle chart of the present invention.
Fig. 2 illustrates the inventive method process flow diagram.
Fig. 3 illustrates city rail speed limit schematic diagram;
Fig. 4, Fig. 5 illustrate the fastest working time of MTDA algorithm schematic diagram;
Fig. 6 illustrates train intelligent control algorithm of the present invention and Yi Zhuang, actual Beijing line ATO effect contrast figure;
Fig. 7 illustrates the effect schematic diagram of method of the present invention under the condition that changes working time;
Embodiment
In order to be illustrated more clearly in the present invention, below in conjunction with preferred embodiments and drawings, the present invention is described further.Parts similar in accompanying drawing represent with identical Reference numeral.It will be appreciated by those skilled in the art that specifically described content is illustrative and nonrestrictive below, should not limit the scope of the invention with this.
The present invention, by simulation driver driving, has set up train intelligent driving expert system and on-line learning algorithm, designs a kind of real-time intelligent control method so that train operation is more steadily comfortable, energy-conservation punctual.This intelligent control method mainly comprises following link: knowledge base, inference machine and gradient descent method, as shown in Figure 1.Wherein knowledge base and inference machine have formed the expert system of simulation driver driving, knowledge base has comprised driving experience, inference machine can according to knowledge base and current train status infer train next step should in operating condition, what expert system drew is the initial output of controller, add the controller amount trimmed that gradient descent method calculates, the final output of synthetic this intelligent controller.
Be illustrated in figure 2 concrete described method flow diagram provided by the invention, the method comprises the steps.
English character variable declaration:
Train real-time status information:
S train position
V train speed
T Train Schedule
Δ t train receives the minimum interval of information
The output of a controller
Information off-line:
(S
1_b, S
2_blS; V
l_1, V
l_2l 0) circuit speed limit
Other English characters and variable:
T
atime precision estimated value
T
etime estimation value
T
cinterval remaining runtime
T
athe maximum time trueness error of ° setting
velocity estimation value
Δ a
maxacceleration maximum changing value
train is in the shortest time of current speed limit
train is in the shortest time of residue distance
the Grad of gradient descent method
λ on-line study step-length
Step 1: at urban track traffic CBTC (Communication based train control system, based on the train automatic controlling system of radio communication) in, the train of operation can be controlled by row the wireless information of central transmission, every time interval through Δ t, obtain train current location s and present speed v information, and the speed limit (S of circuit
1_b, S
2_blS; V
l_1, V
l_2l0), expect the interval S of running time T and Train Stopping
p.If judgement s>=S
p, turn to step 4, enter shutdown phase, otherwise, turn to step 2.As shown in Figure 3.
Step 2: judge the current residing speed restrictive block of train, i.e. S
k-1≤ s≤S
k, k represents the current residing speed restrictive block of train, corresponding speed limit size is V
k.
Step 3: first obtain the general expression of train current working according to following expert knowledge library, ensure comfort level, energy consumption and the punctuality rate of train operation, afterwards according to speed limit situation and the current speed of train, calculate the shortest working time of each speed limit section and remaining runtime is pro rata distributed in each speed limit section, and then drawing the remaining runtime in current speed limit section; Again according to the following inference machine operation information current to train, mainly comprise the working time of present speed, current location, current speed limit, current residual etc., according to knowledge base, infer train after this should in accurate operating condition accelerate, deceleration or coasting.
Expert knowledge library partial content
Knowledge base has mainly comprised some and has ensured the empirical rule that train smooth moves, such as change working rule, as shown in table 1.
Table 1: change working rule
In addition, driver's experience of summary also has:
1, slowly start, acceleration is less (is generally less than 0.6m/s
2);
2, comfortable for ensureing, the each variation of acceleration should be not excessive, and have one period of retention time;
3, the coasting of should trying one's best when high speed, avoids drawing frequently and braking;
4, in strict accordance with car controlling working time requiring: time error is greater than certain error, and train suitably accelerates; Time error within certain error, train coasting.Except employing one's time to the best advantage, more energy-conservation like this;
5, poor according to present speed and speed limit after this, operating mode is handled in judgement in advance, to avoid exceeding ATP speed limit;
In the end accurate shutdown phase, can judge in advance and make corresponding manipulation, with the parking error of avoiding system delay to bring.
Inference machine:
1, the speed limit not declining when train front and when stopping, according to driving experience, the operating mode of train can only be to accelerate or coasting, now again computing time precision estimation value
wherein T
afor time precision estimated value, T
efor time estimation value, T
cfor the remaining runtime in current speed limit section.Just can infer acceleration or coasting operating mode by comparing time precision value and velocity amplitude: if T
a>T
a° and
accelerate wherein T
a° for set maximum time trueness error, T
a°=0.05 be that time error all meets the demands in 5 percent, the size of acceleration is a=a
c+ Δ a
max, the controller that wherein a is next step is initially exported, a
cfor the output of back controller, Δ a
maxfor the maximal value Δ a of acceleration change
max=0.2; Otherwise coasting, coasting Time Controller is finally output as a=0.Note: the remaining runtime T in current speed limit section
c, time Estimate value T
ewith velocity estimation value
derivation algorithm as follows, i.e. shortest time derivation algorithm (Minimal Time Distribution Algorithm, MTDA).
In the time that there is the speed limit of decline in train front, according to driving experience, simultaneously with reference to ATP curve generation method, in certain distance threshold doseag, slow down to prevent from exceeding speed limit.Apart from threshold doseag value be
wherein V
l_nextfor next section speed limit, γ is the number percent that velocity coefficient expects to decelerate to next speed limit, to prevent from exceeding speed limit, and γ=0.9, v is current speed, a
setfor the reference deceleration degree arranging, the size of value should meet comfort level requirement, can not excessive a
set=-0.5m/s
2.After within the scope of train operation enters into apart from threshold doseag, slowing down according to the retarded velocity of setting is that controller is initially output as a=a
set=-0.5m/s
2.Step 2 and the step 3 Train Control Technology based on expert system.
The step of shortest time derivation algorithm (MTDA) proposed by the invention is:
Step1: receive in real time train status information, comprise train speed v, train position s and Train Schedule t, and speed-limiting messages;
Step2: according to train current state (seeing Fig. 4 and Fig. 5), draw train and accelerate to train speed and reach the curve of speed limit with maximum acceleration from train current location, draw the curve of train with maximum braking at the terminal of speed restrictive block simultaneously, as shown in Figure 4, until train is reached home, this curve becomes shortest time curve, represents that train reaches the fastest operation curve of terminal in current location.Can calculate the shortest time of train in current speed limit according to this curve
with the shortest time of train in residue distance
as follows
Step3: calculate the remaining runtime T in current speed limit section
c:
Velocity estimation value
Time Estimate value T
e:
Step4: export in real time T
c,
t
e.
Step 4: train intelligent driving controller is carried out to on-line study, obtain this controller output valve.First the objective optimization function of setting gradient descent method is f=(V
e-v)
2, velocity estimation value again
s
2_bfor current speed limit section end position is the position that speed limit changes, s is the distance of having moved, T
cfor the remaining runtime in current speed limit section;
t
l=T-t, wherein T
lfor remaining runtime, T is that interval requires the overall operation time, and t is the time of having moved; S=∫ vdT=V
avgt, V
avgfor the average velocity moving before this, t is the time of having moved; V=∫ adT=a
avgt, wherein a
avgfor the average acceleration of moving before this.Above-mentioned expansion is brought in objective optimization function, and t is differentiated and show that target function gradient is
According to gradient descent method principle, next step is finally output as to draw controller thus
wherein a
nextfor controller, next step is finally exported, and a is that controller is initially exported, and λ is step-length, and numerical value is λ=0.01,
for Grad.
Step 5: in current city rail friendship system, station platform all can be arranged transponder, facilitates driver to stop according to the locating information of transponder, and the positional information that the present invention provides according to transponder is adjusted described controller output valve.In the time that train need to stop, for meeting precision parking requirement, need utilize the positional information that provides of transponder that station is arranged to adjust several times to reach accurate parking, often adjust controller output during through a transponder.The retarded velocity estimated value of estimating during through i transponder is
wherein V
cfor current speed, S
ibe the distance of i transponder apart from parking terminal, by a
e_irunning to the initial output of i+1 the controller between transponder as train from i transponder is that this Time Controller is initially output as a=a
e_i.The like, to the last stop.In example, this paper method is provided with four transponders, is respectively apart from stop: 102m, and 58m, 13m and 6m, these numerical value derive from the transponder pavingization of certain actual track.
Below in conjunction with embodiment, the present invention is further illustrated.
For convenience of explanation, designed a kind of interval speed limit situation in this embodiment, as shown in Figure 2, this speed limit situation is also circuit speed limit situation the most general between municipal rail train two stations, in figure, and V
l_1, V
l_2and V
l_3be three speed limits, S
1_b, S
2_band S
3_bfor the position that speed limit changes, wherein S
3_bfor stop.We are according to actual track data, and having designed three speed limits is V
l_1=15.99m/s, V
l_2=22.22m/s, V
l_3=15.68m/s and three speed limit change locations are S
1_b=128.54m, S
2_b=1018.4m, S
3_bthe speed limit situation of=1288.9m will be introduced the present invention taking it as circuit case below, and method of the present invention is online real-time, and train is realized from a station runs to the process of the next stop.
P1 (startup): the first speed limit of initialization circuit and speed limit change point, train brings into operation from starting point.According to expert knowledge library, start to start with relatively little acceleration, Train Control device output acceleration is 0.6m/s
2;
P2 (acceleration of speed restrictive block 1): after train brings into operation with Acceleration of starting degree, start real time position, velocity information that Cong Liekong center and mobile unit are received, train is positioned at first speed restrictive block, i.e. 0≤s≤S
1_b(step 1).Due to the front speed limit (step 2 that do not decline, as can be seen from the figure, the speed limit of subordinate phase is greater than the speed limit of first stage), according to the MTDA algorithm in step 3, utilize formula (1)-(5) to distribute in real time remaining runtime, and the optimum output that machine calculates current controller by inference; Because excess time is abundant, this stage continues to accelerate;
P3 (inertia in speed restrictive block 1): train accelerates to and approaches speed limit, according to the expert knowledge library of step 3, train need to slow down to ensure that train speed can not trigger ATP protection, train starts inertia at this moment;
P4 (operation in speed restrictive block 2): train inertia is to next speed restrictive block, because train speed is much smaller than current speed limit, when train operation is apart from s≤S
2_b, in the very long distance in train front, do not decline speed limit and deceleration, described in inference machine rule 1, utilize MTDA formula (1)-(5), computing time estimated value
if T
a>T
a° and V
c<V
e, accelerating, controller is initially exported a=a
c+ Δ a
max, otherwise train coasting, controller is initially exported a=0.Can from simulation result, find out, first train has carried out accelerating to adjust at this one-phase, afterwards because speed is higher, starts inertia.
P5 (braking in speed restrictive block 2): due to V
1_b<V
2_b, there is decline speed limit in train front, described in inference rule 2, calculates range estimation value
as also off-duty S of train
ein distance range time, carry out car controlling according to the method for P4, computing time estimated value
if T
a>T
a° and V
c<V
e, accelerating, controller is initially exported a=a
c+ Δ a
max, otherwise train coasting, controller is initially exported a=0; When train operation is to S
ein distance range time, according to the retarded velocity a setting
set=-0.5m/s
2till decelerating to train and rolling this speed limit section away from, this Time Controller is initially output as a=a
set=-0.5m/s
2.。
P6 (shutdown phase in speed restrictive block 3): when train operation is apart from S
2_b<s≤S
3_bbe that train front needs inlet parking, described in step 5, when train also off-duty to first transponder location during namely apart from the position of stop 102m, while arriving first transponder for guarantee, train speed can not be excessive, so need to decelerate to a desired speed value 10m/s, expect that train arrives first transponder place with the speed of 10m/s, the initial output of this Time Controller
after train arrives last 102m, after this every through a transponder, just according to
calculate estimation retarded velocity, wherein S
1=102, S
2=58, S
3=13, S
4=6, controller is initially exported a=a
e_i.
Note: on the basis of the initial output of controller obtaining at P1-P5, add the amount trimmed of gradient descent method
the final output of synthetic controller
in formula, λ is step-length, and numerical value is λ=0.01,
for Grad, the above-mentioned gradient descent method link of its value computing reference is introduced.
P7: final controller output is applied to train, completes Train Control, and repeat above-mentioned steps in real time, finally complete the even running of train in whole interval.
Be more than the example that we utilize method of the present invention to do train driving, in addition, we also utilize the measured data of Yi Zhuang, Beijing line, and ATO (pid control algorithm control) data contrast with the result of this method.That Fig. 6 shows is the travelling speed curve comparison figure of intelligent method of the present invention and PID method, and wherein, be 100.2 seconds expectation working time that the inventive method is set, and final actual run time is 100.7 seconds, within error range.Table 2 is main performance index contrasts, can find out that the inventive method is obviously more comfortable energy-conservation, and meet the index of time and stopping accuracy.Meanwhile, the present invention also has larger adaptability working time to expecting, we have designed three and have expected working time: 95.2 seconds, 100.2 seconds and 110.2 seconds.What Fig. 7 showed be three expects the travelling speed comparison diagram of working time, and table 3 is main performance index contrasts separately.Can find out, the present invention has larger dirigibility to requiring working time, has broken through traditional deficiency that is controlled at this respect.In summary, the inventive method can be simulated driver driving experience substantially, has greatly improved existing control method, improve comfort level, reduced energy consumption, met stopping accuracy requirement, meanwhile, have larger dirigibility working time to expecting, this has also made up the deficiency of existing control method.
Table 2
| Contrast index | PID method | Intelligent method of the present invention |
| Actual run time (s) | 101.2 | 100.7 |
| Stopping accuracy (cm) | -1 | 5.3 |
| Comfort level | 11.34 | 7.22 |
| Unit mass energy consumption (J) | 209.7 | 152.7 |
Note: the less expression of comfort level value is impacted less, thereby more comfortable (lower same).
Table 3
| Contrast index | Expected time 95.2s | Expected time 100.2s | Expected time 110.2s |
| Actual run time (s) | 96.5 | 100.7 | 112.8 |
| Stopping accuracy (cm) | 5.8 | 5.3 | 4.3 |
| Comfort level | 7.78 | 7.22 | 7.15 |
| Unit mass energy consumption (J) | 170.8 | 152.7 | 114.9 |
To sum up, the train automatic Pilot (ATO) that technical solution of the present invention is different from the past, utilizes the real time data in train travelling process, by simulation driver's driving strategy, do not need off-line rate curve and train model parameter accurately, realize train intelligent driving.The present invention, from a brand-new angle, has experience driver's driving strategy by simulation, utilize the real time data of train operation, by reasoning and conclusion, and the output of controlled device; Utilize real-time information and the speed-limiting messages of train, by reasonable distribution excess time of train operation, obtain time precision estimated value and velocity estimation value; Consider the speed limit of circuit to the impact of train driving, if the speed limit of next stage is greater than the speed limit when front section, utilize expert inferential mechanism and time precision estimated value, velocity estimation value to obtain current Train Control device defeated; MTDA algorithm based on real-time information, utilizes train operation speed limit and current location, speed and time, and derivation train operation is expected working time and expected travelling speed; Train intelligent driving proposed by the invention considers needs punctuality rate, overspeed protection, comfortableness and the train operation energy consumption taken into account in train driving process, ensure the many factors in train driving process as far as possible; And the present invention proposes the train intelligent driving on-line study method based on gradient method, define objective function, and by asking for the gradient of objective function, derived on-line study formula; Set up the train on-line intelligence driving technology of complete simulation driver driving, and utilize the actual operation data of subway Yi Zhuang, Beijing line, with the contrast of train on-line intelligence driving simulation result, result has shown that the method, in saving energy consumption, improves the superiority of the aspects such as punctuality rate.
Obviously; the above embodiment of the present invention is only for example of the present invention is clearly described; and be not the restriction to embodiments of the present invention; for those of ordinary skill in the field; can also make other changes in different forms on the basis of the above description; here cannot give all embodiments exhaustively, everyly belong to apparent variation or the still row in protection scope of the present invention of variation that technical scheme of the present invention extends out.
Claims (5)
1. the urban railway transit train intelligent control method based on driving experience and on-line study, is characterized in that, this control method comprises the steps:
1), when train operation, every the time interval of Δ t, determine train current location s and present speed v information, and the speed limit (S of circuit
1_b, S
2_blS; V
l_1, V
l_2l 0), expect the interval S of running time T and Train Stopping
pif, judgement s>=S
p, turn to step 5, enter shutdown phase, otherwise, turn to step 2;
2) judge the current residing speed restrictive block of train, i.e. S
k-1≤ s≤S
k, k represents the current residing speed restrictive block of train, corresponding speed limit size is V
k;
3) obtain the current operating mode of train according to expert knowledge library, again according to speed limit situation and train present speed, calculate the shortest working time of each speed limit section and remaining runtime is pro rata distributed in each speed limit section, and then drawing the remaining runtime in current speed limit section; Again by inference machine to the current operation information of train infer after this whether train accelerate, deceleration or coasting;
4) train intelligent driving controller is carried out to on-line study, obtain this controller output valve;
5) train operation circuit is arranged to transponder, and the positional information providing according to transponder is adjusted described controller output valve.
2. a kind of urban railway transit train intelligent control method based on driving experience and on-line study according to claim 1, it is characterized in that, in described step 3, the current operation information of train comprises present speed, current location, current speed limit and current residual working time.
3. a kind of urban railway transit train intelligent control method based on driving experience and on-line study according to claim 1, it is characterized in that, in described step 3, it is theorized that machine to the current operation information of train infer after this whether train accelerate, deceleration or coasting specifically comprise:
The speed limit not declining when train front and when stopping, the operating mode of train can only be to accelerate or coasting, now again computing time precision estimation value
wherein T
afor time precision estimated value, T
efor time estimation value, T
cfor the remaining runtime in current speed limit section, just can infer acceleration or coasting operating mode by comparing time precision value and velocity amplitude: if T
a>T
a° and
accelerate wherein T
a° for set maximum time trueness error, the size of acceleration is a=a
c+ Δ a
max, the controller that wherein a is next step is initially exported, a
cfor the output of back controller, Δ a
maxfor the maximal value Δ a of acceleration change
max=0.2; Otherwise coasting, coasting Time Controller is finally output as a=0;
In the time that there is the speed limit of decline in train front, according to ATP curve generation method, in certain distance threshold doseag, slow down to prevent from exceeding speed limit, apart from threshold doseag value be
wherein V
l_nextfor next section speed limit, γ is the number percent that velocity coefficient expects to decelerate to next speed limit, to prevent from exceeding speed limit, and γ=0.9, v is current speed, a
setfor the reference deceleration degree arranging, the size of value should meet comfort level requirement, can not excessive a
set=-0.5m/s
2, within the scope of train operation enters into apart from threshold doseag after, according to the retarded velocity of setting, to slow down be that controller is initially output as a=a
set=-0.5m/s
2.
4. a kind of urban railway transit train intelligent control method based on driving experience and on-line study according to claim 1, is characterized in that, described step 4 is carried out on-line study to train intelligent driving controller and comprised:
1) the objective optimization function of setting gradient descent method is f=(V
e-v)
2, be input as the current operation information of train;
2) judge the current residing speed restrictive block of train, i.e. S
k-1≤ s≤S
k, k represents the current residing speed restrictive block of train, corresponding speed limit size is V
k;
3) according to following process computation velocity estimation value
and V
e:
T
L＝T-t；
4) train operation distance and train present speed can be with representing as follows:
s＝∫vdT＝V
avgt；
v＝∫adT＝a
avgt；
Wherein, V
avgfor the average velocity moving before this, a
avgfor the average acceleration of moving before this;
5) bring above formula into objective optimization function, and to the t gradient that obtains objective function of differentiating be:
6), based on gradient descent method, known Train Control device output is adjusted into
λ is step-length, and numerical value is λ=0.01,
for Grad.
5. a kind of urban railway transit train intelligent control method based on driving experience and on-line study according to claim 1, it is characterized in that, when the positional information providing according to transponder in described step 5 is adjusted described controller output valve, the retarded velocity estimated value of estimating during through i transponder is
wherein V
cfor current speed, S
ibe the distance of i transponder apart from parking terminal, by a
e_irunning to the initial output of i+1 the controller between transponder as train from i transponder is that this Time Controller is initially output as a=a
e_i, the like, to the last stop.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410283677.0A CN104134378A (en) | 2014-06-23 | 2014-06-23 | Urban rail train intelligent control method based on driving experience and online study |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410283677.0A CN104134378A (en) | 2014-06-23 | 2014-06-23 | Urban rail train intelligent control method based on driving experience and online study |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104134378A true CN104134378A (en) | 2014-11-05 |
Family
ID=51807036
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410283677.0A Pending CN104134378A (en) | 2014-06-23 | 2014-06-23 | Urban rail train intelligent control method based on driving experience and online study |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104134378A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104464482A (en) * | 2014-12-12 | 2015-03-25 | 柳州铁道职业技术学院 | Vehicle-mounted ATC-MMI test method for urban rail transit |
| CN104554299A (en) * | 2014-11-21 | 2015-04-29 | 北京交控科技有限公司 | Automatic train drive method based on ATP/TD loop line mode |
| CN104564914A (en) * | 2015-01-16 | 2015-04-29 | 太原矿机电气股份有限公司 | Intelligent control system for equipment train |
| CN104881527A (en) * | 2015-05-13 | 2015-09-02 | 南京理工大学 | ATO (Automatic Train Operation) speed command optimization method of urban rail transit train |
| CN105785795A (en) * | 2016-05-05 | 2016-07-20 | 北京交通大学 | Train operation speed curve energy saving optimization method based on particle swarm algorithm |
| CN106569499A (en) * | 2016-11-11 | 2017-04-19 | 华南智能机器人创新研究院 | Robot adaptive control servo driving method and system |
| CN106844947A (en) * | 2017-01-18 | 2017-06-13 | 清华大学 | A kind of locomotive energy saving optimizing automatic Pilot method based on high-order relational learning |
| CN105551337B (en) * | 2015-12-21 | 2018-02-09 | 北京交通大学 | A kind of train operator's auxiliary driving method and system |
| CN107948570A (en) * | 2017-11-29 | 2018-04-20 | 长沙安行机电设备有限公司 | A kind of rail track real scene video plays synchronous locomotive driving real time information source method |
| WO2018121691A1 (en) * | 2016-12-29 | 2018-07-05 | 比亚迪股份有限公司 | Train automatic control method and apparatus, and vehicle-mounted automatic train operation system |
| CN109204390A (en) * | 2018-09-29 | 2019-01-15 | 交控科技股份有限公司 | A kind of Train control method based on deep learning |
| CN109445304A (en) * | 2018-09-25 | 2019-03-08 | 交控科技股份有限公司 | A kind of intelligent fault analysis system and method based on cab signal |
| CN109858154A (en) * | 2019-01-31 | 2019-06-07 | 广州地铁设计研究院股份有限公司 | A kind of energy-saving train operation method based on multiple-objection optimization |
| CN110490367A (en) * | 2019-07-15 | 2019-11-22 | 西安理工大学 | Bullet train automatic Pilot energy conservation optimizing method based on maximal principle |
| CN110663073A (en) * | 2017-06-02 | 2020-01-07 | 本田技研工业株式会社 | Policy generation device and vehicle |
| CN111376949A (en) * | 2018-12-29 | 2020-07-07 | 交控科技股份有限公司 | Method for calculating adjustment speed of ATO quasi-point curve |
| CN112249096A (en) * | 2020-09-14 | 2021-01-22 | 南京铁道职业技术学院 | Accurate parking method for urban rail transit station |
| CN112286045A (en) * | 2020-10-15 | 2021-01-29 | 天津津航计算技术研究所 | Start-stop control system for high-speed train with vacuum metal pipeline |
| CN112722014A (en) * | 2019-10-14 | 2021-04-30 | 中车唐山机车车辆有限公司 | Railway vehicle overspeed control method and device and terminal equipment |
| CN112977537A (en) * | 2021-04-02 | 2021-06-18 | 湖南中车时代通信信号有限公司 | Vehicle speed control method and device and related equipment |
| CN113562020A (en) * | 2021-07-22 | 2021-10-29 | 北京交通大学 | High-speed train parking control method and system simulating human intelligence |
| CN113640380A (en) * | 2021-06-07 | 2021-11-12 | 济南大学 | Multi-stage classification method and system for rail damage detection |
| CN115366946A (en) * | 2022-08-04 | 2022-11-22 | 王学浩 | Rail vehicle positioning method, device, equipment and storage medium |
| WO2023134093A1 (en) * | 2022-01-14 | 2023-07-20 | 福建工程学院 | Virtual subway train operation big data generation method based on interval speed limit and expert experience |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1967155A (en) * | 2006-11-01 | 2007-05-23 | 上海轨道交通设备发展有限公司 | Error comprehensive diagnosis method and device for mass transit device maintenance system |
| CN101941450A (en) * | 2010-08-26 | 2011-01-12 | 北京交通大学 | Method and system for controlling working condition switching of train |
| CN101992782A (en) * | 2010-12-02 | 2011-03-30 | 张婧 | Urban unmanned electronic railcar taxi traffic system |
| US20120004796A1 (en) * | 2010-04-01 | 2012-01-05 | Alstom Transport Sa | Method for managing the circulation of vehicles on a railway network and related system |
| CN102442323A (en) * | 2011-10-25 | 2012-05-09 | 中国电子科技集团公司第三十二研究所 | Method for achieve automatic driving curve generation between stations during operation by automatic train driving system |
| CN102514591A (en) * | 2011-12-27 | 2012-06-27 | 北京交通大学 | Method for achieving train parking at stations through train braking ratio dynamic regulation |
| CN102514602A (en) * | 2011-12-29 | 2012-06-27 | 浙江网新中控创新技术研究开发有限公司 | Method and system for planning and controlling train travelling speed |
| CN102981408A (en) * | 2012-12-10 | 2013-03-20 | 华东交通大学 | Running process modeling and adaptive control method for motor train unit |
| CN103019267A (en) * | 2012-12-10 | 2013-04-03 | 华东交通大学 | Predicative control method for modeling and running speed of adaptive network-based fuzzy inference system (ANFIS) of high-speed train |
| CN103092076A (en) * | 2013-01-06 | 2013-05-08 | 华东交通大学 | Multi-model adaptive pictorial information digitizer (PID) control of braking process of motor train unit |
| CN103246200A (en) * | 2013-04-17 | 2013-08-14 | 华东交通大学 | Synchronous tracking and controlling method for motor train unit based on distributed model |
| CN103777631A (en) * | 2013-12-16 | 2014-05-07 | 北京交控科技有限公司 | Automatic driving control system and method |
-
2014
- 2014-06-23 CN CN201410283677.0A patent/CN104134378A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1967155A (en) * | 2006-11-01 | 2007-05-23 | 上海轨道交通设备发展有限公司 | Error comprehensive diagnosis method and device for mass transit device maintenance system |
| US20120004796A1 (en) * | 2010-04-01 | 2012-01-05 | Alstom Transport Sa | Method for managing the circulation of vehicles on a railway network and related system |
| CN101941450A (en) * | 2010-08-26 | 2011-01-12 | 北京交通大学 | Method and system for controlling working condition switching of train |
| CN101992782A (en) * | 2010-12-02 | 2011-03-30 | 张婧 | Urban unmanned electronic railcar taxi traffic system |
| CN102442323A (en) * | 2011-10-25 | 2012-05-09 | 中国电子科技集团公司第三十二研究所 | Method for achieve automatic driving curve generation between stations during operation by automatic train driving system |
| CN102514591A (en) * | 2011-12-27 | 2012-06-27 | 北京交通大学 | Method for achieving train parking at stations through train braking ratio dynamic regulation |
| CN102514602A (en) * | 2011-12-29 | 2012-06-27 | 浙江网新中控创新技术研究开发有限公司 | Method and system for planning and controlling train travelling speed |
| CN102981408A (en) * | 2012-12-10 | 2013-03-20 | 华东交通大学 | Running process modeling and adaptive control method for motor train unit |
| CN103019267A (en) * | 2012-12-10 | 2013-04-03 | 华东交通大学 | Predicative control method for modeling and running speed of adaptive network-based fuzzy inference system (ANFIS) of high-speed train |
| CN103092076A (en) * | 2013-01-06 | 2013-05-08 | 华东交通大学 | Multi-model adaptive pictorial information digitizer (PID) control of braking process of motor train unit |
| CN103246200A (en) * | 2013-04-17 | 2013-08-14 | 华东交通大学 | Synchronous tracking and controlling method for motor train unit based on distributed model |
| CN103777631A (en) * | 2013-12-16 | 2014-05-07 | 北京交控科技有限公司 | Automatic driving control system and method |
Non-Patent Citations (3)
| Title |
|---|
| 冷勇林，陈德旺，阴佳腾: "基于专家系统及在线调整的列车智能驾驶算法", 《铁道学报》 * |
| 刘炯: "基于专家系统的城轨CBTC列车运行调整研究", 《中国优秀硕士学位论文全文数据库》 * |
| 张强等: "城市轨道交通ATO系统性能指标评价", 《都市快轨交通》 * |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104554299B (en) * | 2014-11-21 | 2017-05-31 | 交控科技股份有限公司 | Train automatic Pilot method based on ATP/TD loop wire standards |
| CN104554299A (en) * | 2014-11-21 | 2015-04-29 | 北京交控科技有限公司 | Automatic train drive method based on ATP/TD loop line mode |
| CN104464482A (en) * | 2014-12-12 | 2015-03-25 | 柳州铁道职业技术学院 | Vehicle-mounted ATC-MMI test method for urban rail transit |
| CN104564914A (en) * | 2015-01-16 | 2015-04-29 | 太原矿机电气股份有限公司 | Intelligent control system for equipment train |
| CN104564914B (en) * | 2015-01-16 | 2016-05-25 | 太原矿机电气股份有限公司 | A kind of equipment train intelligence control system |
| CN104881527A (en) * | 2015-05-13 | 2015-09-02 | 南京理工大学 | ATO (Automatic Train Operation) speed command optimization method of urban rail transit train |
| CN105551337B (en) * | 2015-12-21 | 2018-02-09 | 北京交通大学 | A kind of train operator's auxiliary driving method and system |
| CN105785795A (en) * | 2016-05-05 | 2016-07-20 | 北京交通大学 | Train operation speed curve energy saving optimization method based on particle swarm algorithm |
| CN105785795B (en) * | 2016-05-05 | 2018-11-23 | 北京交通大学 | A kind of train running speed curve energy conservation optimizing method based on particle swarm algorithm |
| CN106569499A (en) * | 2016-11-11 | 2017-04-19 | 华南智能机器人创新研究院 | Robot adaptive control servo driving method and system |
| WO2018121691A1 (en) * | 2016-12-29 | 2018-07-05 | 比亚迪股份有限公司 | Train automatic control method and apparatus, and vehicle-mounted automatic train operation system |
| CN106844947A (en) * | 2017-01-18 | 2017-06-13 | 清华大学 | A kind of locomotive energy saving optimizing automatic Pilot method based on high-order relational learning |
| CN106844947B (en) * | 2017-01-18 | 2019-11-26 | 清华大学 | A kind of locomotive energy saving optimizing automatic Pilot method based on high-order relational learning |
| CN110663073B (en) * | 2017-06-02 | 2022-02-11 | 本田技研工业株式会社 | Policy generation device and vehicle |
| CN110663073A (en) * | 2017-06-02 | 2020-01-07 | 本田技研工业株式会社 | Policy generation device and vehicle |
| CN107948570A (en) * | 2017-11-29 | 2018-04-20 | 长沙安行机电设备有限公司 | A kind of rail track real scene video plays synchronous locomotive driving real time information source method |
| CN107948570B (en) * | 2017-11-29 | 2020-09-04 | 长沙安行机电设备有限公司 | Real-time information source method for real-scene video playing and synchronous locomotive running of railway line |
| CN109445304A (en) * | 2018-09-25 | 2019-03-08 | 交控科技股份有限公司 | A kind of intelligent fault analysis system and method based on cab signal |
| CN109204390A (en) * | 2018-09-29 | 2019-01-15 | 交控科技股份有限公司 | A kind of Train control method based on deep learning |
| CN109204390B (en) * | 2018-09-29 | 2021-03-12 | 交控科技股份有限公司 | Train control method based on deep learning |
| CN111376949A (en) * | 2018-12-29 | 2020-07-07 | 交控科技股份有限公司 | Method for calculating adjustment speed of ATO quasi-point curve |
| CN111376949B (en) * | 2018-12-29 | 2022-02-15 | 交控科技股份有限公司 | Method for calculating adjustment speed of ATO quasi-point curve |
| CN109858154A (en) * | 2019-01-31 | 2019-06-07 | 广州地铁设计研究院股份有限公司 | A kind of energy-saving train operation method based on multiple-objection optimization |
| CN110490367A (en) * | 2019-07-15 | 2019-11-22 | 西安理工大学 | Bullet train automatic Pilot energy conservation optimizing method based on maximal principle |
| CN112722014A (en) * | 2019-10-14 | 2021-04-30 | 中车唐山机车车辆有限公司 | Railway vehicle overspeed control method and device and terminal equipment |
| CN112722014B (en) * | 2019-10-14 | 2022-11-08 | 中车唐山机车车辆有限公司 | Railway vehicle overspeed control method and device and terminal equipment |
| CN112249096A (en) * | 2020-09-14 | 2021-01-22 | 南京铁道职业技术学院 | Accurate parking method for urban rail transit station |
| CN112286045A (en) * | 2020-10-15 | 2021-01-29 | 天津津航计算技术研究所 | Start-stop control system for high-speed train with vacuum metal pipeline |
| CN112977537A (en) * | 2021-04-02 | 2021-06-18 | 湖南中车时代通信信号有限公司 | Vehicle speed control method and device and related equipment |
| WO2022205847A1 (en) * | 2021-04-02 | 2022-10-06 | 湖南中车时代通信信号有限公司 | Vehicle speed control method and apparatus, and related device |
| CN113640380A (en) * | 2021-06-07 | 2021-11-12 | 济南大学 | Multi-stage classification method and system for rail damage detection |
| CN113640380B (en) * | 2021-06-07 | 2023-05-09 | 济南大学 | Multistage classification method and system for rail damage detection |
| CN113562020A (en) * | 2021-07-22 | 2021-10-29 | 北京交通大学 | High-speed train parking control method and system simulating human intelligence |
| WO2023134093A1 (en) * | 2022-01-14 | 2023-07-20 | 福建工程学院 | Virtual subway train operation big data generation method based on interval speed limit and expert experience |
| CN115366946A (en) * | 2022-08-04 | 2022-11-22 | 王学浩 | Rail vehicle positioning method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104134378A (en) | Urban rail train intelligent control method based on driving experience and online study | |
| CN103235843B (en) | A kind of urban railway transit train running optimizatin control simulation method and system | |
| CN102981408B (en) | Running process modeling and adaptive control method for motor train unit | |
| Bai et al. | Distributed cooperative cruise control of multiple high-speed trains under a state-dependent information transmission topology | |
| CN108475465B (en) | Method for moving, in particular controlling or regulating a vehicle fleet | |
| Bertoni et al. | An adaptive cruise control for connected energy-saving electric vehicles | |
| CN104228887B (en) | The method and system that a kind of ATO grade is adjusted in real time | |
| CN105551337A (en) | Driving auxiliary method and system for train driver | |
| CN102514602A (en) | Method and system for planning and controlling train travelling speed | |
| CN104881527A (en) | ATO (Automatic Train Operation) speed command optimization method of urban rail transit train | |
| CN104590333A (en) | Railway train intelligent operation optimization control system | |
| CN108538069A (en) | A kind of ring road merging area car speed managing and control system and method | |
| CN102951189B (en) | Based on the high speed train of safe distance between vehicles real-time calibration with progress control method of speeding | |
| CN103761371B (en) | Vehicle follow gallop system self-adaption control method based on model reference | |
| CN105035090A (en) | Trace prediction control method for autonomously driven vehicle based on traffic signal lamp | |
| CN103863364A (en) | Freight locomotive automatic-control real-time optimization control system based on dispatch signal | |
| CN109634281B (en) | Train system modeling and distributed cooperative control method | |
| CN109760721A (en) | A kind of train interval operation real-time regulating system and method | |
| CN109693688A (en) | The train automatic driving control system and control method of default capabilities | |
| Pan et al. | Dynamic control of high-speed train following operation | |
| CN114834503A (en) | Virtual marshalling train control method based on elastic tracking model | |
| Rahn et al. | Energy-efficient driving in the context of a communications-based train control system (CBTC) | |
| Zhou et al. | The safety braking protection model of virtually coupled train platoon in subway | |
| Huang et al. | Cooperative tracking control of the multiple-high-speed trains system using a tunable artificial potential function | |
| Zhou et al. | Model predictive control for high-speed train with automatic trajectory configuration and tractive force optimization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141105 |
|
| WD01 | Invention patent application deemed withdrawn after publication |