Methodological Aspects for the Development of Information Systems

of Unmanned Mobile Vehicles

Sergey Sokolov

and Andrey Boguslavsky

Keldysh Institute of Applied Mathematics, Miusskaya Sq. 4, Moscow, Russia

Keywords: Unmanned Mobile Vehicles, Information Support Systems, Real Time Vision Systems, Programming

Framework, Interpreting Navigation, Configuration Space.

Abstract: In the work there is analyzed a process of developing and arranging information support systems of mobile

vehicles with the enhanced autonomous level and fully autonomous. There are reviewed current trends in

creating this sort of systems. In order to increase the effectiveness of processes of creating intellectual

autopilots it is proposed to introduce into the review the space of similar systems configuration. The space

axes are: sensors, models of environment, and algorithmic support. The points of this space are “assemblage

points” of particular application systems. The configuration space enables to efficiently resolve the process

of creating information support systems with the help of compliance matrixes, evaluate possible options of

arranging the particular system. There are proposed solutions on components per each axis of the

configuration space. As the basis for mobile means information systems the real time multifunctional vision

systems and software-hardware unification is proposed. In terms of the hardware components, there are

reviewed, in detail, arrangements of real time vision systems. The construction of external world models is

proposed to establish on the interpreting navigation concept. In terms of algorithmic support, the

combination of the split-level processing algorithms is proposed to implement based on large-scale frames

and unification of software implementations. Examples of using the above-described approaches and

solutions are presented.

1 INTRODUCTION

The peculiarity of robotics is integrity, integration of

achievements of various fields of knowledge. The

replacement of a person in his/her activities in the

world around us requires the perception of this

world and skill to move intentionally within it. The

information support is an essential part of mobile

and actually the main part of autonomous robots.

The feature of robotics on the present stage is a shift

from the purely scientific disciplines and narrowly

specialized spheres of application into the more

demandable means of reviewing technological and

economical aspects of creating robotics systems

(RS). The prevailing issues are reliability and

economic efficiency of the systems being created. It

is possible to determine two principle trends of

creating this sort of systems. The first one (we guess,

the dominating one) is the development/provision of

needed transport infrastructure in which there is a

motion of autonomous/automatic devices. The

second trend is the intellectualization of autopilots.

The main distinguishing feature is the effort to

overcome the infrastructure disadvantages or

complexities of external conditions at the expense of

increasing capabilities of the on-board analyzer.

Up to now there are known a few projects of

creating mobile robotics systems with the enhanced

autonomous level. After the competitions in terms of

purposeful motions of autonomous ground vehicles

under the aegis of the US Defense Advanced

Research Projects Agency (DAPRA) (Buehler et al,

2007, 2009) there have been activated works on

creating commercial versions of similar systems.

The well-known corporation Google since 2010 has

been working on the project of “self-managed” light

vehicle (Markoff, 2010). The US automobile

company Oshkosh Truck Corporation has developed

based on the heavy hauler a vehicle-robot (Defence

Talk, 2006). The prototype of this vehicle

participated in races DARPA 2004 and 2005. The

US automobile company Ford demonstrated an

autonomous Ford on the snow – driving in hard

viewing conditions (on the practice ground in

Michigan) (Ford Motor Company, 2016). Nissan

492

Sokolov, S. and Boguslavsky, A.

Methodological Aspects for the Development of Information Systems of Unmanned Mobile Vehicles.

DOI: 10.5220/0006003904920498

In Proceedings of the 13th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2016) - Volume 2, pages 492-498

ISBN: 978-989-758-198-4

(North Korea) based on the standard electric car

Nissan Leaf is developing an autonomous vehicle

which is passing field tests (Kadakov, 2015).

Cognitive Technologies (Russia) in September 2015

notified about tests of autonomous KAMAZ on the

practice ground in Naberezhnye Chelny (informed

by Cognitive Technologies press-service), in

October 2015 there appeared massages about tests of

unpiloted KAMAZ on the practice ground in

Noginsk (KAMAZ, 2015).

As examples of the most interesting non-

commercial projects of unmanned mobile vehicles

it’s possible to name two US projects with more than

ten-year history. NASA program of creating and

using MARS rovers (1997 – present time) (Ellery,

2016). DAPRA program (USA) of creating

unmanned flight vehicles, in particular, X47B

project (lenta.ru, 2012).

The solution of tasks of information support of

intellectual mobile RS requires the incorporation of

scientific-technical achievements of such areas as

applied mathematics, sensory, computer science,

mechanics, control theory, artificial intellect, etc.

About the complexity of tasks set in front of the

developers of similar systems say the resources and

the time dedicated by the developers to meet the

target goals. The total cost of NASA and Jet

Propulsion Lab project aimed at creating the MARS

rover Curiosity Rover 2012 is estimated as 2.5bln

USD. Such powerful companies as Google and Ford

having more than a ten-year study experience join

their efforts for the creation of unmanned vehicle

(Automotive News, 2015). The Russian program of

creating the unmanned vehicle is estimated as

300bln RUB. ROSAVTODOR is already spending

annually 4bln RUB on the road marking for the

unmanned vehicles (Alizar, 2015).

2 CONFIGURATION SPACE

In order to find way around a multidimensional

space of problems when solving particular practical

tasks of intellectual mobile robotics and determine

the ways of forming technologies of development,

we propose to structure the variety of the

components of the information support systems

(ISS) within control systems over such objects and

make a review in the configuration space. Each axis

of this space represents a plurality of possible

solutions/arrangements of respective elements.

Along one axis, there are sensors, along the second

one there are view models of the environment, along

the third one there are information processing

algorithms. In the conventional cube of this space,

it’s possible to select alternative ways of collecting

and processing information for the intellectual

management over the intentional motions of mobile

robots.

In the review we’ll pay special attention to the

unification of the software-hardware components as

an important means of supplying reliability and

economic efficiency of the proposed solutions.

2.1 Sensor Support of the Onboard ISS

The common list of means of sensor support systems

of control over autonomous mobile objects at

present is rather clearly identified. Shortly it may be

characterized as follows. Traditional internal sensors

of the mobile means plus sensors of external world.

For the ground mobile vehicles such sensors are

radars (as a rule, LIDARs), gyroscopes,

accelerometers, satellite navigation systems and

vision system (VS). VS is the one, “external”

component of the information support system of the

ground unmanned vehicles, which represents the

principle interest of our review. For the unmanned

vehicles in other spaces (air, underwater) this set

kind of changes with account of environment

properties (for example, ultra-sound detectors and

side-scan sonars for the underwater vehicles,

phased-array radars, radio and barometric altimeters

for flight vehicles), but does not change the system-

wide software-hardware ISS architecture of

intellectual control systems. As a note, the

characteristics of applied sensors are closely

connected with requirements to the capacities of the

onboard computers providing acquisition and

processing of the respective data.

Analysis of the current state of the developments

in the sphere of mobile vehicles ISS enables to

identify such trends in the arrangement of mobile

vehicles with sensor systems.

The priority issues of arrangement are those of

complexation: all sensor systems; sensor systems of

various ranges of radiant energy; 2D and 3D data.

VS of the visual, IR, radar bands in totality or

separately become an essential component of

intellectual autopilots.

We’ll discuss in more detail the requirements to

the most widespread and demandable VS type, i.e.

visible range VS. We should note such specific

requirements to video-cameras as part of VS of the

contemporary mobile vehicles as:

– resolution (from 2Mb and higher);

– random access to raster units;

– signal output to the unified digital channel,

Methodological Aspects for the Development of Information Systems of Unmanned Mobile Vehicles

493

possible in various modes, but with obligatory

mode “no compression”;

– responsivity/dynamic range;

– automatic control over the optical system

stopping;

– external synchronization.

In order to assure requirements to the

hardware/sensor devices when arranging the mobile

vehicles ISS there is widely used COTS technology,

i.e. method is well-known, well-proven for the

arrangement of new highly-sophisticated

complicated systems. In terms of sensory provision

as means of gathering information about external

world in real-time scale, there are well-proven 3D

selective control detectors, in combination with

visual data about the same spatial domain. As

examples of such solutions we can indicate both

which have become well-known Lidar devices (for

“outdoor” application”), Kinect (for “indoor”

application”), or similar, and author’s developments

of VS with operated structured illumination based on

spherical motors of direct control (Sokolov et al,

1995).

Universal computer based on IBM PC

compatible architecture is the basis for integration of

module solutions. We propose a computing control

part of the real-time VS as the most resource-

intensive, to take as the basis of arrangement of the

computing part of the information support system of

the intellectual control systems. Article 4 will

provide examples of successful application of such

approach.

2.2 Presentation of the External World

in the Autopilot Information

Support System

The problem of describing the external world in the

computer presentation is not new and there is a

range of approaches to its solution. The use of

external world model in the onboard computer of

mobile facility imposes additional constraints both

on the computer capacities and on the real-time scale

where it’s necessary to operate the model.

Notwithstanding various studies in the area of

external world simulation, we should state the

absence of databases to store object-oriented data

with the access in real-time scale.

Presentation about the external world in the

onboard control systems of mobile vehicles is

constructed with the use of the so called “navigation

cross” (Figure 1).

When solving navigation tasks we take as the

basis for the construction of external world model

the interpretive navigation concept that has been

developing since 80ies last century (Sokolov and

Kirilchenko, 2015). The basis of the concept is made

of following statements. As the presentation of

knowledge about the external world there is used not

a quantitative model in the absolute coordinate

system, but a qualitative one as a sequence of

interchange areas with same informative-visual

contents. The model is formed as the information

equivalence graph (IEG) This presentation serves a

supplement and development of approaches of

SLAM in the provision of goal-oriented motions of

mobile vehicles in a weakly structured, unfamiliar

environment.

GRNSS





MOBILE

VEHICLE

CONTROL

SYSTEM



CNS



OPERATOR

Figure 1: Diagram “navigation cross”, where GRNSS –

global radio-navigation satellite system; CNS –

conventional navigation system including processes of

reckoning and adjustments by reference points; IN –

interpretive navigation.

The arrangement of the external world model as

IEG is cost-effective with regard to the computer

capacities and enables to unify the motion planning

stage independently from the level of highlighting

the reference marks in the surrounding environment.

In the software implementation of the external

world presentation there is used a unified instrument

tool to work with the graphs. This structure

implements polymorphism during the work with

various structures. It uniformly operates with feature

descriptions of the landmarks, circles and belts of

the landmarks, areas of the information equivalence,

anthologies of the textual description of scenes and

sets of visual data processing algorithms.

The movement control language in the mode of

tele-programming based on interpretive navigation is

more natural for the communication of the operator

and unmanned transport vehicle as it applies “the

commonly-accepted” logics of explaining the route

based on dynamic pattern in the process of

movement along this route of visible descriptions of

the environment based on the reference marks.

The issues of unified presentation of the

environment model are on the stage of elaboration

and reaching the technological solutions.

ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics

494

2.3 Algorithmic Supporting

When reviewing the main issues of the vision

system in the current intellectual autopilots and

determining the main requirements to them, there

are defined the most specific groups of algorithms of

detecting and identifying objects on images. For the

criteria of the algorithms quality, it is proposed to

review robustness, localization, and computing

implementation, in particular, with such a

characteristic as the number of operations per pixel

in-process. The effectiveness of performing these

requirements depends on the selected technology of

receiving and preliminary processing of the input

information about the observed object space, applied

structure solutions and algorithms of transforming

information showing the observed object space into

the applied model. All this in totality determines the

main characteristics of the system, such as

persistence, validity of recognition, accuracy of

performed actions, and technical feasibility.

There are determined ways of satisfying

requirements to the algorithmic supporting of the

intellectual mobile vehicles: rational data selection;

effective processing; hardware implementation;

sensory systems integration of ISS; unification of

software solutions.

The above ways are shown in the software-

hardware architecture. In particular, in order to

implement low and intermediate algorithm levels the

following hardware facilities are widely applied:

Multi-core general processors (Sokolov and

Boguslavsk, 2011); GPU (Vasilyev et al., 2011);

other special video-processors; FPGA;

Combination of the above-listed (Krasnobaev, 2005)

Review of the available publications and

conferences’ documents shows that some computer

vision tasks have been brought to the stage of

technological solutions, i.e. known combination of

algorithms of visual data collection and processing.

Results of the recent software surveys (Laplante,

2004; Rankin al., 2011; Hillel et al., 2012; Tumofte

et al., 2014) enabled to activate and extend the VS

application in practical robotics tasks. There are

works proposing metrics for the quality evaluation

of the real time vision systems (Baranov and

Telezhkin, 2014; Baranov and Nikiforov, 2015).

Practical possibility of using stereo-systems in real

time scale has been justified (Wagner et al., 2002).

The above achievements in the area of

algorithmic supporting leave open a question about

the technology of complexing the algorithmic

supporting of particular application tasks.

2.4 Proposed Implementation of

Algorithmic Supporting

Distinctive features of the proposed approach to the

formation of the software for the information

support of unpiloted mobile vehicles are as follows:

• Provision of possibility for the cross-platform

development based on universal PC and high-

speed carry to the special computing platforms at

the expense of division of the software into the

set of interacting parallel subsystems.

• Specialization of the original VS programming

frame at the expense of extension within the

developed software architecture for the

interaction with external subsystems of

unmanned vehicles.

• The extended subsystem of processing visual

data from a few point of view assuring the real-

time processing of video-sequences from high-

resolution visual sensors.

• Use of reusable components to process the visual

data for the prototyping of software being

developed.

• Implementation of special debugging tools to

assure the reproducibility of the software

operation on the development and testing stage.

Fig.2 shows unified structural units of the VS frame.

Subsystem for the

visual data input

Start

Client registration

Turn on

Notification after the image capturing

Open the image buffer

Copy image

Close the image buffer

Get the current

Image

Process image

View the current

processing results

Turn off

Client registration cancelling

Subsystem for

processing images

“The VS adjustment

mode” object

“Frame buffer”

object

GUI Subsystem

Figure 2: Unified structural units of the VS frame: а) –

system setup mode; b) – automatic mode.

Implementation of all algorithms included into

Methodological Aspects for the Development of Information Systems of Unmanned Mobile Vehicles

495

the software frame are supplied with characteristics

of performance time on some or other hardware

support tools.

Direct analogs for the software as functional

frame and set of typical modules for unmanned and

semiautomatic vehicles are not available. Some

solutions from related subject area are known.

• Functional libraries implementing algorithms of

processing images and computer vision

(OpenCV, LeadTools, HALCON) do not contain

full applications and implementations of the

application level architectures.

• Rapid application development environment to

process images (Matrox Assistant, National

Instruments LabVIEW) are aimed at industrial

automation tasks and use of closed software-

hardware platform. They are not available on the

level of initial texts and do not enable to control

the functioning of the application in real-time

mode.

• Prototypes of algorithms (extension packs

MATLAB) are aimed at adjusting some

algorithms which require the transfer or repeated

implementation for the use in the target

application.

The described tools are good on the stage of the very

first surveys, but during the shift to the technologies

of implementation from these surveys to the onboard

means there is a big distance that may have hard

obstacles such as real-time scale requirements and

coherence of all components.

In comparison with all above-described

solutions, the proposed software system simplifies

the design and incorporation of information systems

based on visual information processing at the

expense of using the worked-out software

components and expended software model (Sokolov

and Boguslavsky, 2011).

2.5 ISS Arrangement Scheme

Here-below there is an approximate mobile means

ISS arrangement scheme of the intellectual mobile

robot to solve a specific task.

The arrangement process is iterational.

1. Wording of the formalized description of a task

in the language of object description, external

world (target environment) and necessary actions

(transformations) with the objects in its state

space. (The efficient approach of this

arrangement stage is a proven method of

constructing anthologies of the subject field).

2. The external world model is being formed.

3. There are selected hardware sensory tools

assuring the receiving of needed initial data (to

form the external world model).

4. There is formed a set of algorithms for the

information support of solving the goal-oriented

task.

5. Testing of capability of finding/reaching all

above-listed components and calculation of the

possession cost.

6. Definition: arrangement is acceptable? If yes, the

process is completed; if not, the transfer to the

next iteration (steps 1-5)

For a quantitative estimation of MV ISS

configuration process quality it is possible to use

deviation size (Δ) of current ISS configurations

from system of requirements / the technical project.

This size is defined by the function of the

coordination depending on matrixes of conformity in

planes of configuration space. In the described initial

stage of researches linear function of the

coordination was considered Δ = X - W Σα

where: X - a set (tuple) of variables of the

configuration space meeting requirements to packed

ISS; R

- a conformity matrix; α

- importance factor;

W - a matrix of the dimension coordination.

3 EXAMPLES OF

IMPLEMENTATIONS

We’ll enlist a few examples of using the above-

described unified software in tasks of mobile

vehicles information support.

Control over railway infrastructure objects.

Railway infrastructure objects are quite different and

need a regular state control. Vision systems are

widely used in the mobile control facilities. They

help to collect data, and the results of control are

formed in the deferred analysis. We proposed

intellectual mobile modules which enable to process

the visual data «on line” (Sokolov et al., 2012). The

above-described approaches permitted to shortly

solve a few control tasks of such railway

infrastructures as position finding of the contact wire

heightwise and on the map, determination of

interacting force of the contact wire and pantograph;

control over realigning of rails, inspection over the

shoulder of ballast section and other additional

parameters of the track structure. All measurements

and control data are fixed to absolute (GIS and

GPS/GLONASS) and relative (route) coordinates.

Implementation of the algorithmic supply on

three general-purpose computers with Intel Pentium

4 2.4 GHz processor enables to inspect objects in

ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics

496

one field of view at speed up to 250 km/h.

Automatic monitoring surveillance system of

approximation and docking processes of space crafts

and ISS. One of the most important stages of the

space craft flight is their approximation and

coupling. Based on the visual data there has been

developed Automatic monitoring surveillance

system of approximation and coupling processes of

space crafts and ISS. The system enables to

determine, in automatic mode, parameters of relative

movement of space crafts and ISS, analyzes their

compliance with the permitted values and shows

information in a way acceptable for the perception

by the person-operator or automatic control system

(Boguslavsky et al., 2004).

Night vision system to help the driver. VS based

on thermal camera analyzes environment in front of

the moving vehicle and provides information about

the vehicle position with regard to roadway, warns

about possible traffic obstacles such as other

vehicles or pedestrians. The system mock-up

showed good results both in urban environment and

on country roads (Sokolov et al., 2008).

The implementation of algorithmic supply on

one general-purpose computer with Intel Core Quad

2.4 GHz processor enables to acquire appropriate

information at speed up to 90km/h.

VS as part of mobile laboratory for online

diagnostics of road surface. VS to control the state

of the toad surface is intended to check wheel

tracking, cracks, holes, defects of road marking. VS

consists of looking-forward video-camera and three

cameras with structured lighting. All results of

visual data processing are integrated with

GPS/GLONASS data about the positioning of

mobile laboratory and electronic route map (Sokolov

et al., 2012).

Implementation of the algorithmic supply on

three general-purpose computers with Intel Core i7

2.4 GHz processors enables to inspect the road

surface at speed up to 120 km/h.

VS as part of mobile complex of operational

mapping. VS task as part of mobile complex of

operational mapping is putting, on the electronic

map, objects located along the route of the mobile

complex. Besides VS, which consists of two stereo-

pairs, the complex includes GIRS, satellite

navigation system, odometer, and program making

e-maps. Data from all the subsystems are integrated

for the precise location of the objects found within

view. Some objects are identified and recorded into

GIS in automatic mode (road signs and other known

engineering facilities), some other part is fixed in

semiautomatic mode with the operator’s

involvement (Sokolov et al., 2011).

The implementation of algorithmic supply on

three general-purpose computers with Intel Core i7

1.8 GHz processors enables to make operational

mapping at speed up to 60km/h.

Unmanned flying vehicle VS as part of the

onboard navigation equipment for piloted and

unmanned flying vehicles. Based on the VS

development of the ground mobile vehicles there is

formed a VS software-hardware architecture for

information support of automatic movement of a

flying vehicle on the runway, take-off and landing.

On the VS mock-up there have been obtained

assessments of accuracy characteristics of the

system. Experiments with the mock-up as part of the

flying laboratory confirmed a possibility of

successful application of the vision system as part of

supplementary independent information channel for

the provision of automatic control over the flying

vehicle when moving along the RW, take-off and

landing (Sokolov et al., 2015).

4 CONCLUSION

There have been reviewed trends in creating

information support systems of mobile vehicles with

the enhanced level of autonomous and fully

autonomous. It is stated a shift of mobile robotics

from the rank of purely scientific disciplines and

narrowly-specialized fields of application into the

rank of more demandable tool of solving a wide

range of tasks of human activity. Vision systems

occupy the first roles in systems of information

support of mobile vehicles. In order to increase the

efficiency of processes of creating intellectual

autopilots of the mobile vehicles and to reduce a

price of the autopilots it is proposed to introduce for

the review a configuration space of such systems. In

this space, it is proposed a methodology of arranging

the information support systems of the mobile

vehicles with the real-time vision systems as central

part.

On the basis of generalization of requirements to

hardware and unification of VS software

architecture and its realization in the form of

software framework the universal architecture of

systems of autopilots information support is

generated. Examples of economically expedient

solution of applied problems of land, air and space

application with use of the described technique are

resulted. In the meantime, a wide range of questions

still needs their settlement. This is also a

construction of more distinct metrics on the axis of

the proposed configuration space and construction of

Methodological Aspects for the Development of Information Systems of Unmanned Mobile Vehicles

497

matrix of the components of information support

systems compliance, and some others. The solution

of the specified problems makes the program of the

further works of our group.

ACKNOWLEDGEMENTS

The work partly involves grants RFBR № 15-08-

06341, 16-08-01282 and grant RSF № 16-19-10705.

REFERENCES

Buehler, M. et al., 2007. The 2005 DARPA Grand

Challenge, Springer.

Buehler, M. et al., 2009. The DARPA Urban Challenge,

Springer.

Markoff, J., 2010. Google Cars Drive Themselves, in

Traffic. In The New York Times. October 11, 2010.

Self-driving car, https://www.google.com/selfdrivingcar

Defence Talk, 2006. https://lenta.ru/news/2006/01/26/

Oshkosh.

Ford Motor Company, 2016. http://www.ford.ru/

AboutFord/News/Vehicle/2016/Vhcl_160113.

Kadakov, М., 2015. Calculate cameras of unmanned

Nissan Leaf.http://www.zr.ru/content/articles/839794 -

derzhim-ruki-u-rulya-samoupravlyaemogo-nissan-leaf.

KAMAZ Company, 2015. http://www.kamaz.ru/press/

releases/bespilotnyy_kamaz_uzhe_na_doroge.

Ellery, A., 2016. Rover vision – fundamentals. In

Planetary Rovers, Springer, pp.199-262.

Lenta.ru, 2012. USA have finished deck tests unmanned

X-47B. http://lenta.ru/news/2012/12/20/x47b/.

Automotive News. Ford and Google companies will unite

efforts on unmanned cars. https://www.drive.ru/news

/ford/567936dc95a65602f50000c2.html.

Alizar A. Pilotless "Kamazy" leave on general purpose

roads. Geektimes, https://geektimes.ru/post/264602/.

Sokolov, S.M., Max, D.P., Wallace, R.S., 1995. Simple

multi function vision system for 3D data acquisition.

In Proceedings of ICAR95, 7th International

Conference on Advanced Robotics. Barcelona, Spain,

vol. 2, pp. 674-678.

Sokolov, S.M., Kirilchenko, A.A., 2015. Information

support of autonomous land vehicles on the basis of

interpreting navigation. In News SFU. Engineering

science. №1, p. 120-133.

Sokolov S.M., Boguslavsky A.A. Intellectual Images

Processing for a Realtime Recognition Problem. //

Proc. The 2nd Intern. Multi-Conf. on Complexity,

Informatics and Cybernetics (IMCIC2011), Orlando,

Florida, USA, March 27

-30

, 2011, Orlando, Florida,

USA, Vol. II, pg.406-411.

Vasilyev, A.I., Boguslavskiy, A.A., Sokolov, S.M., 2011.

Parallel SIFT-detector implementation for images

matching. In Proc. of the 21st Conference on

Computer Graphics and Vision, GraphiCon’2011.

September 26-30, 2011, Moscow, pp. 173-176.

Krasnobaev A.A. The review of detecting an image simple

elements algorithms and the analysis of their hardware

realization of possibility. Proc. KIAM RAS. 2005, http:

//keldysh.ru/papers/2005/prep114/prep2005114.html

Laplante, Software engineering for image processing

systems. CRC Press, 2004.

Rankin al., Unmanned ground vehicle perception using

thermal infrared cameras // Proc. SPIE Unmanned

Systems Technology XIII, 2011(SPIE, 2011).

Hillel A., Lerner R., Levi D., Raz G. Recent progress in

road and lane detection: a survey // Machine Vision

and Applications, February 2012.

Tumofte R., Zimmermann K., Van Gool L. Multi-view

traffic sign detection, recognition, and 3D localisation

// Machine Vision and Applications, April 2014,

Vol.25(3), pp.633-647.

Baranov, S.N., Nikiforov V.V. Density of Multi-Task

Real-Time Applications // Conference of Open

Innovation Association, FRUCT, 2015-June, pp. 9-15.

Baranov S.N., Telezhkin A.M. Metrics for Software

Development. // SPIIRAS Proceedings, issue 5(36),

2014. p. 5-27.

Wagner M. et al. Principles of Computer System Design

for Stereo Perception // Carnegie-Mellon University

Technical Report, CMU-RI-TR-02-01, 2002.

Sokolov, S.M., Boguslavsky, A.A., Trifonov, O.V.,

Computer vision as a part of systems of intellectual

control on railway transportation. In Proc. of the first

scientific and technical conference «Intellectual

control systems on a railway transportation».

November,15-16,2012,Moscow,OpenSociety "NIAS".

Boguslavsky, A.A., Sazonov, V.V., Sokolov, S.M.,

Smirnov, A.I., Saigiraev, K.U., 2004. Automatic

Vision-based Monitoring of the Spacecraft Docking

Approach with the International Space Station. In Proc

of the ICINCO 2004, First International Conference

on Informatics in Control, Automation and Robotics.

Setúbal, Portugal, Vol. 2, pp. 79-86.

Sokolov, S.M., Boguslavsky, A.A., Platonov, A.K., Kiy,

K.I., Gorelik, L.I., Filachev, А.М., Fumin, А.I., 2008.

An IR Channel-Based Automated Driver Assistance

System. In Proc. WMSCI’08, 13th Intern. Conf on

Systemics, Cybernetics and Informatics. Orlando,

Florida, USA, July 29-Aug 3, 2008, vol.III, p.368-373.

Sokolov, S.M., Boguslavsky, A.A., Vasilyev, A.I.,

Trifonov, O.V., 2012. Methodical and practical

aspects of system engineering of computer vision for

navigation and measurements. In News SFU.

Engineering science. №3, pp.76-85.

Sokolov, S.M., Boguslavsky, A.A., Vasilyev, A.I.,

Trifonov, O.V, Nazarov, V.G, Frolov, R.S., 2011. A

mobile complex for operative creation and updating of

navigation maps. In News SFU. Engineering science.

№3, pp. 157-166.

Sokolov, С.М., Boguslavsky, А.А., Feudorov, N.G.,

Vinogradov P.V. Vision System for information

support of automatic landing and movement on

runway. In News SFU. Engineering science, 2015,

№1, pp. 96-109.

ICINCO 2016 - 13th International Conference on Informatics in Control, Automation and Robotics

498