AU2021100003A4 - A deep transportation model to predict the human mobility for autonomous vehicle - Google Patents
A deep transportation model to predict the human mobility for autonomous vehicle Download PDFInfo
- Publication number
- AU2021100003A4 AU2021100003A4 AU2021100003A AU2021100003A AU2021100003A4 AU 2021100003 A4 AU2021100003 A4 AU 2021100003A4 AU 2021100003 A AU2021100003 A AU 2021100003A AU 2021100003 A AU2021100003 A AU 2021100003A AU 2021100003 A4 AU2021100003 A4 AU 2021100003A4
- Authority
- AU
- Australia
- Prior art keywords
- deep
- human
- transportation
- predict
- model
- 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.)
- Ceased
Links
- 230000033001 locomotion Effects 0.000 claims abstract description 16
- 238000000034 method Methods 0.000 claims abstract description 15
- 238000013135 deep learning Methods 0.000 claims abstract description 11
- 238000013508 migration Methods 0.000 claims abstract description 5
- 230000005012 migration Effects 0.000 claims abstract description 5
- 238000012549 training Methods 0.000 claims abstract description 5
- 238000004088 simulation Methods 0.000 claims description 9
- 230000015654 memory Effects 0.000 claims description 6
- 230000006399 behavior Effects 0.000 claims description 5
- 238000013461 design Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000007619 statistical method Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 238000012800 visualization Methods 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims 1
- 230000001934 delay Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 claims 1
- 238000011161 development Methods 0.000 abstract description 5
- 230000006855 networking Effects 0.000 abstract description 3
- 238000011156 evaluation Methods 0.000 abstract description 2
- 238000011160 research Methods 0.000 abstract description 2
- 230000018109 developmental process Effects 0.000 abstract 2
- 238000013473 artificial intelligence Methods 0.000 abstract 1
- 230000001149 cognitive effect Effects 0.000 abstract 1
- 230000003247 decreasing effect Effects 0.000 abstract 1
- 238000001514 detection method Methods 0.000 abstract 1
- 230000007613 environmental effect Effects 0.000 abstract 1
- 239000000446 fuel Substances 0.000 abstract 1
- 238000004458 analytical method Methods 0.000 description 6
- 210000004027 cell Anatomy 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000005094 computer simulation Methods 0.000 description 2
- 239000012141 concentrate Substances 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000001994 activation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000009529 body temperature measurement Methods 0.000 description 1
- 230000009172 bursting Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013079 data visualisation Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000011273 social behavior Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
- G08G1/012—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from other sources than vehicle or roadside beacons, e.g. mobile networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
Abstract
A DEEP TRANSPORTATION MODEL TO PREDICT THE HUMAN MOBILITY FOR
AUTONOMOUS VEHICLE
ABSTRACT:
With significant demographic expansion and urbanization, transport overcrowding has been a huge
significant crisis globally. Urban sprawl triggers major socioeconomic losses annually worldwide
associated with fuel consumption, unnecessary environmental damage, loss of opportunity and
decreased productivity. Recognizing how people travel and choose method of transportation via a
large-scale transportation system is a key to traffic congestions detection and traffic management.
Human mobility is a diverse field in physics and computer science and has gained a great deal of
interest in recent decades. Such descriptive structures and predictive methods have been
formulated for modeling and assessing human movement. Even so, multi-source interdependent
information like portable devices, GPS and social networking sites ensure a innovative driving
factor for analyzing metropolitan trends of human movement. More and more exponential
development has been made across the last decade in self-driving automated vehicles, primarily
powered by developments in deep learning and artificial intelligence. This invention acquires
broad and diverse information and constructs a cognitive model called Deep transportation model,
to simulate and forecast human mobility in the autonomous vehicles. The core feature of deep
transportation model is the deep learning framework like LSTM architecture that strives to
recognize the aspects of human mobility and transport from broad and streaming data. The deep
transportation model dynamically visualize or accurately predict migration of people and their
means of transportation in a large-scale transportation system depending on the training paradigm,
provided any time span, particular area of the city or observable migration of people. Extensive
research and evaluations show the reliability and outstanding quality of our model and indicate
that human modes of transport can be projected and interpreted more effectively.
1
A DEEP TRANSPORTATION MODEL TO PREDICT THE HUMAN MOBILITY FOR
AUTONOMOUS VEHICLE
Drawings:
Input
SPre-pocessin ModuleDeep _earIng Moduk
Visuaiztion and Evalatio
L--- .- - - - - - -.- .- - - - - .-.-------2
-4
Figure 1: The architecture of a deep transportation model to predict the human
mobility for autonomous vehicle.
1
Description
Drawings:
Input
ModuleDeep _earIng Moduk Visuaiztion and Evalatio SPre-pocessin
L--- .- - - - - - -.- 2 .- - - - - .-.-------
-4
Figure 1: The architecture of a deep transportation model to predict the human mobility for autonomous vehicle.
Description
Field of the Invention:
This invention relates to develop a deep transportation model for predicting the human mobility of autonomous vehicle. Multi-task deep learning methodology assures that the algorithm can train from broad sequence information at various time spans and create a prediction model. The deep transportation model provides certain metropolitan environment, span of time or demonstrated human movement and accurately forecast or visualizes a high proportion of people's behavior and their traffic migration.
Background of the invention:
Transport system performs a key responsibility in community and the environment while at the same moment, contributes greatly to reduce emissions. In current history, automated vehicle systems have been set in motion as portion of the intelligent and innovative mobility strategy. There is a greatest opportunity that completely independent on-road vehicles will produce a range of human and economic opportunities afforded that AVs concentrate on a renewable launch vehicle.
A variety of experiments are being performed in the last generations simulation and recognizing transport infrastructure. These models based primarily on small-scale transportation systems or concentrate mostly on mathematical models or data visualization. Even so, analysis on the complexities of human movement and the development of travel on a nation-wide or city-wide level is very constrained owing to the unavailability of a consistent methodology to the detailed detecting of human mobility.
According to Aschenbruck, a detailed analysis of explicitly obtainable projected path data sources is provided. The examples illustrate a few obstacles in the analysis of trajectory data, including such information fusion, data interpretability, Affirmation, and meteorological conditions.
Treurniet et al suggest a categorization of miniature mobility applications in the field of cell networks. The categorization relies attention on the subsequent subsections, like spatial constraints, goal selection, destructible terrain, acceleration, queue length, and social behavior.
Barbosa et al explore various techniques and simulations which analyze dynamics of human movement from a computer vision viewpoint and categorize modern practice into three categories. Though this is the distinction from such a review is which we are not just to concentrating on machine learning, and moreover evaluating the three alternative analytical techniques of human traffic density.
In accordance with Hess the basic phases of prototype development and evaluation from an objective standpoint are described. This section presented valuable knowledge for researchers who wish to test their conceptual simulation methods. However, these insights are restricted to the realm of electronic networking in digital environments and has to be modified in vehicular environments.
Currently, with the and proliferation of positioning technology, human wearable sensor information including GPS monitoring of mobile devices, CDR data and online social networking data are now big data that provides the capability to consider human mobility and urban transit circumstances at the city-wide stage. In addition, awareness, simulation and extraction of human mobility and their means of transportation have been the key subject of study for smart urban management and effective urbanization. Fortunately, many of these experiments are focused on limited datasets, and the simulations described are indeed shallow models which encounter experience in balancing a broad and diverse sample collection.
According to statistical analyses Toch, that the design is developed to describe significant variations of versatility and to establish a myriad of features. However, similar research on the estimation of urban transport is still not exposed. The current Reviews have been published advice on diverse features of human migration, so there is no prior in-depth analysis of issues and opportunities to data-oriented analysis and simulation in vehicular environments.
In relation to the forecast of vehicular traffic or acceleration by J. Sun, and also being reported to forecast accidents and achieved. In this study, they allowed it to forecast Collapses in the roads.
They exploited the vehicle movement information which have been classified into various groups including reflect congestion, free flow and flow of jam. Using this Strategy, they expected the traffic accidents community, whereas they registered high precision, splitting acceleration variables into various predicted value and it does not quite reflect the consistency of the estimated parameters. Perhaps it could be increased while using appropriate data for limited screens to show a far more realistic view of the actual state of the lane.
In some other works, developers have employed a variety of several other methods, other than deep learning techniques, to forecast travel demand at bridges or to simulate transit time. While several observers have recorded very better prognostic outcomes by obtaining better precision and reduced error rates, and there is a necessity to investigate in detail in this future to ensure reliability and to allow use of such a new innovation to realize the importance of transportation control. Often the direction of vehicles and other independent variables may be employed to forecast traffic requirements on transport networks, however regardless of congestion provides the duration the driver pulls when traveling the range in a unit of period, so that it would be more necessary to understand vehicular pressures.
While an in-depth interpretation of the interdependent multi-source information can offer a rigorous comprehension of the human being transportation societies. Deep learning technology has progressively become proven to be a significant successful method of learning which has shown outstanding success in a number of fields like vision, voice, text, and transportation. This invention thus represents first ever attempt to highlight a profound instructional strategy to simulating human displacement and public transportation.
Objects of the Invention:
• The main objective is to develop a deep transportation model to predict the human mobility for autonomous vehicle. • The second is to gather massive and multiethnic information and design an abstract model for forecasting and simulate human movement and travel on a substantial transportation network.
* The Learning architecture is capable of dynamically training human movement and vehicle transfer methods from a stratified dataset and enables the network is trained through training samples at various timespans.
Summary of the Invention:
The proposed framework comprises of four primary modules: a database server, a preprocessing module, a deep learning module, and a simulation and measurement module. The database server module collects and preserves the origins of the information. It will offer facilities for archiving, extraction, formatting, and visualisation. The pre-processing module will clean up the details and link person movement to the transit system. Finally, this interacts to produce a vast range of individual GPS tracks with a transit service mark on large-scale transportation system. Deep learning framework is a significant aspect of Deep Transport and contains four LSTM layers for training: one encoding layer for distinct feedback series, one decoding layer for distinct output series, and the other two layers are computational models which express the same specifications. The visualisation and assessment module will eventually simulate the outcomes and assess the reliability of the integrated method.
Human movement and travel behaviors have a significant level of spatiotemporal association. Based on the spatial characteristics of society movement, the Recurrent Neural Networks (RNNs) are well suited for capturing the spatiotemporal development of humanoid travel and travel transfer behaviors. A standard LSTM neuron that comprises an input gate, a forget gate ft, a cell ct, an output gate ot, and an output response ht. The input gate and the forget gate control the transmission of data into and out of the unit. The output gate determines how much information is transmitted from the neuron to the outgoing ht. The memory location has a self-connected recurrent weight 1 edge, meaning that the differential will travel through several effective integration until disappearing or bursting. It thus overwhelms the complications in the learning of the RNN model induced by the vanishing gradient consequence.
The pervasive implementations of smartphones and portable technologies prompt to an abundance of multi-source information linked to human movement, offering an innovative and performance characteristics of urban human mobility trends. These databases are obtained implicitly, like call information (CDR), credit card, smart card, the object of which is not to gather mobility data but to record all purchases. But these multi-source interdependent databases document people's consumption movements and propose possible trends of movement.
Detailed Description of the Invention:
Figure 1: The architecture of Automatic Hand Sanitizer Dispenser with Contactless Temperature Measurement model.
Figure 2: Deep learning architecture.
Figure 3: LSTM memory block.
Detailed Description of the Invention: Figure 1 explains the overall architecture of deep transportation model to predict the human mobility. The input data called heterogeneous data for human mobility is gathered as GPS records and transport data from multiple sources. Retrieve the necessary knowledge from big data and then exploit any or more of its features to extract valuable information. The process is termed data parsing. Then refine datasets attribute values to remove some other valuable knowledge and create innovative features that can be considered as inputs to deep LSTM network. Some of input variables without any alteration in their measurements are employed and analyzed some of them to retrieve any other relevant insights. The timestamp function is utilized to derive the values of hours, minutes, days, months, year and day of the week. Excavation of relevant features is a difficult task that involves the depth knowledge of the sample dataset parameters and their influence on the learning phase. LSTM is equipped to recognize the statistical analysis for a long time frame and continuously evaluate the estimated time latency for forecasting. In this analysis, LSTM is chosen to predict the long-term quantitative dependence of urban environments and traffic behavior.
Figure 2 illustrates the LSTM architecture. Long Short-Term Memory is a type of recurrent neural network (RNN). The output of RNN from the final phase is provided as input in the current time step. It resolved the issue of long-term RNN dependences in which the RNN could not forecast the phrase retained in long-term memory, rather could provide more precise evidence from modern equipment. As the duration of the distance rises, RNN does not have efficient efficiency. By design, LSTM can hold data for a significant period of time. It is intended for encoding, forecasting and categorizing data on the basis of time series. The LSTM has a chain framework that includes four computational models and separate memory structures labeled cells. The data is processed in the cells and the storage deceptions are performed by the gates.
Figure 3 demonstrates the memory block of LSTM. It manipulates a mapping from an input sequence X to an output sequence Y by estimating the network unit activations iteratively.
Claims (6)
1. A deep transportation model to predict the human mobility for autonomous vehicle comprises Database servers Data preprocessing Deep learning module Visualization process.
2. According to claim 1, broad and diverse data sources are utilized to comprehend human movement and their means of transportation. A vast volume of individual traffic activities is employed to develop the public transport model, which provides traffic details such as road links, distance travelled and the intensity of each route for crisis situations.
3. From claim 2, the data servers are employed for storing the acquired GPS data. Data server system includes five computers with specifications including Intel Xeon 2.6 GHz CPU, 8 GB RAM, and 2x2 TB HDD for constructing a Hadoop cluster which comprises of 32 cores, 32 GB memory, and 16 TB storage.
4. From claim 1, the data preprocessing technique is employed to retrieve the necessary knowledge from big data and then exploit any or more of its features to extract valuable information.
5. Based on claim 1, the deep learning module like LSTM is introduced for prediction. LSTM is enabled to recognize the statistical analysis for a long period of time and accurately evaluate the estimated time delays for simulation. LSTM is employed to design the long term perceptual dependence of human migration and travel behaviors.
6. From claim 1, K-fold cross-validation was conducted to test the efficiency. The full sequence data was arbitrarily divided into 3 samples one sample was considered as testing set, whereas the other two were being used as training samples. For the simulation process, the program randomly generated patterns relying on the output transfer in the deep module unit.
A DEEP TRANSPORTATION MODEL TO PREDICT THE HUMAN MOBILITY FOR AUTONOMOUS VEHICLE
Drawings: 2021100003
Figure 1: The architecture of a deep transportation model to predict the human mobility for autonomous vehicle.
Figure 2: Deep learning architecture
Figure 3: LSTM memory block
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2021100003A AU2021100003A4 (en) | 2021-01-01 | 2021-01-01 | A deep transportation model to predict the human mobility for autonomous vehicle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2021100003A AU2021100003A4 (en) | 2021-01-01 | 2021-01-01 | A deep transportation model to predict the human mobility for autonomous vehicle |
Publications (1)
Publication Number | Publication Date |
---|---|
AU2021100003A4 true AU2021100003A4 (en) | 2021-03-18 |
Family
ID=75093487
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
AU2021100003A Ceased AU2021100003A4 (en) | 2021-01-01 | 2021-01-01 | A deep transportation model to predict the human mobility for autonomous vehicle |
Country Status (1)
Country | Link |
---|---|
AU (1) | AU2021100003A4 (en) |
-
2021
- 2021-01-01 AU AU2021100003A patent/AU2021100003A4/en not_active Ceased
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Chen et al. | A hybrid CNN-LSTM model for typhoon formation forecasting | |
Miglani et al. | Deep learning models for traffic flow prediction in autonomous vehicles: A review, solutions, and challenges | |
Park et al. | Interpretation of Bayesian neural networks for predicting the duration of detected incidents | |
CN110570651A (en) | Road network traffic situation prediction method and system based on deep learning | |
Niu et al. | A novel spatio-temporal model for city-scale traffic speed prediction | |
Parsa et al. | A data-driven approach to characterize the impact of connected and autonomous vehicles on traffic flow | |
Olayode et al. | Prediction and modeling of traffic flow of human-driven vehicles at a signalized road intersection using artificial neural network model: A South African road transportation system scenario | |
Yang et al. | Intelligent bridge management via big data knowledge engineering | |
CN114664091A (en) | Early warning method and system based on holiday traffic prediction algorithm | |
Rahmani et al. | Graph neural networks for intelligent transportation systems: A survey | |
CN110555551B (en) | Air quality big data management method and system for smart city | |
CN112801340B (en) | Crowd density prediction method based on multi-level city information unit portraits | |
CN115148019A (en) | Early warning method and system based on holiday congestion prediction algorithm | |
Hamad et al. | Predicting freeway incident duration using machine learning | |
Xiong et al. | A Descriptive Bayesian Approach to Modeling and Calibrating Drivers' En Route Diversion Behavior | |
Xing et al. | A data fusion powered bi-directional long short term memory model for predicting multi-lane short term traffic flow | |
Lei et al. | Prediction of PM2. 5 concentration considering temporal and spatial features: A case study of Fushun, Liaoning Province | |
Feng et al. | A survey of visual analytics in urban area | |
AU2021100003A4 (en) | A deep transportation model to predict the human mobility for autonomous vehicle | |
Liu et al. | Highway traffic congestion detection and evaluation based on deep learning techniques | |
Chuanxia et al. | Machine learning and IoTs for forecasting prediction of smart road traffic flow | |
Chawla et al. | Real-time traffic congestion prediction using big data and machine learning techniques | |
Xiao et al. | Embedding geographic information for anomalous trajectory detection | |
ABBAS | A survey of research into artificial neural networks for crime prediction | |
Saremi et al. | Alternative Sequence Classification of Neural Networks for Bridge Deck Condition Rating |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FGI | Letters patent sealed or granted (innovation patent) | ||
MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |