TABLE2.2: Aggregated preference table
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 Z
C1 x 3 3 3 3 3 3 3 3 3 100
C2 0 x 2 3 3 3 3 3 3 3 76
C3 0 1 x 3 3 3 3 3 3 3 71
C4 0 0 0 x 1 0 0 0 0 0 0
C5 0 0 0 2 x 0 1 0 0 0 11
C6 0 0 0 3 3 x 1 2 0 1 36
C7 0 0 0 3 2 2 x 2 2 1 41
C8 0 0 0 3 3 1 1 x 1 1 36
C9 0 0 0 3 3 3 1 2 x 2 47
C10 0 0 0 3 3 2 2 2 1 x 44
2.5. Example on the Proposed Method 23 A linear transformation f (Z) = a·Z+bwas executed on the scale in order to get unit weight sum. Parametera=0.002 andb =0.01 in this case. Offset (b) has to be defined in order to all the criteria have weights different from 0.
The transpose matrix of trapezoidal hesitant fuzzy decision matrix had the form:
{(0.6,0.7,0.8,0.9)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)}
{(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)}
{(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.2,0.3,0.40.5),(0.4,0.5,0.6,0.7)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.4,0.5,0.6,0.7),(0.8,0.9,1,1)} {(0,0.1,0.2,0.3),(0.2,0.3,0.4,0.5)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0,0.1,0.2,0.3),(0.4,0.5,0.6,0.7),(0.8,0.9,1,1)} {(0.2,0.3,0.40.5),(0.4,0.5,0.6,0.7),(0.6,0.7,0.8,0.9)} {(0,0.1,0.2,0.3),(0.6,0.7,0.8,0.9))}
{(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)} {(0.6,0.7,0.8,0.9),(0.8,0.9,1,1)}
In case of L2 The trapezoidal hesitant fuzzy decision matrix contains the dif- ferent (all the occurring samples are present once) linguistic values (trape- zoidal fuzzy numbers) of linguistic terms formulated by the experts as mea- sure of compliance of the given alternative to the given criterion. For ex- ample: E1,1 contains the values {0.6, 0.7, 0.8, 0.9} as all the experts stated the first alternative (A1 - L2 layout) compliance is ’Fairly high’ in terms of Safety/OSH/fire protection aspects (C1). Example for evaluation of the lay- outs: In terms of C4 (Traveled distance by human labor) the difference is obvious. All the auxiliary facilities with flexible locations (used by human labor) are in one group in the middle of the site in case of L2, so this alterna- tive acquired the highest scores.
Equation2.3yields the matrix of expected values, which transpose matrix is:
0.75 0.8375 0.8375 0.8375 0.8375 0.8375
0.8375 0.65 0.8375
0.8375 0.45 0.65
0.8375 0.65 0.65
0.7417 0.65 0.65
0.8375 0.7375 0.35
0.8375 0.65 0.65
0.5417 0.55 0.45
0.8375 0.8375 0.8375
This matrix aggregates the opinions of all the experts. So, finally con- ventional KIPA-matrix (Table2.3) with preference (c) and disqualificance (d) values can be constructed (based on pair-wise comparison) and preference order can be decided (detailed in [104]), this gave us the following result:
L2>L4>L6.
TABLE2.3: KIPA-matrix
L2 L4 L6
L2 cij [% ] - 69 80 dij [% ] - 9 9 L4 cij [% ] 56 - 84
dij [% ] 14 - 14 L6 cij [% ] 61 81 -
dij [% ] 22 17 -
The winner configuration is the same (the further preference order is not) as in case of the original case study in [141], they state that total handling costs were reduced with this. There is no standard way to evaluate the qual- ity of a decision making method [227], so objective final preference order cannot be achieved either, because the results depends on the experts sub- jective opinion. The authors of [141] generated different feasible layout ver- sions with multi-objective optimization using Pareto-based ant colony opti- mization and single-objective optimization using Max-min ant system. For the multi-objective optimization two objective functions were considered: f1: score for likelihood of accidents happened associated with the site layout, f2: score for total handling cost between the facilities associated with the construction site layout; in case of single-objective optimization the objective
2.6. Conclusion 25
27
Chapter 3
Analysis of Vehicle Fleet Tracking
In the complex process of construction planning, if the layout is designed (Chapter 2), the vehicle fleet is provided [21] and the work and position change routes are carefully planned [22], next we have to monitor our ma- chines during the motion. It is inevitable for safety reasons, dynamic path planning and it is also a requirement in order to be able to install higher level intelligence in our machines, e.g. environment perception. The method in Chapter4 needs point cloud registration, so the vehicle position must be known and the algorithm in Chapter 5 requires the knowledge of the ve- hicle’s ego motion as well. In this Chapter, I present criteria of selecting ap- propriate localization system and make suggestions about different scenarios based on evaluation of different alternatives, earlier there were no such spe- cific criteria system. These results were presented in [108]. The presented results of the thesis (criteria definition) is own contribution. My coauthor in the above mentioned article contributed as supervisor in the publication. He helped me to select the methods and checked the correctness of the experi- mental methods and results.
The aims of using so called vehicle tracking systems (VTS) vary in wide range. The most common ones can be categorized as surveillance, control, navigation, data prediction and customer service. Naturally, because of the high level of development of mobile machine tracking other applications are showing up continuously. Fuel, vehicle security monitoring, transit and as- set tracking the main elements of surveillance category but one can also find here such things as temperature monitoring of food delivery vans. By mea- suring and changing some parameter from distance the remote controlling of the covering motion of unmanned vehicle just as feasible as some ma- chine parts’. These systems can help in field service management and sales by locating the closest expert to a customer or the customers nearby to the salesman location, so the scheduling can be much easier and quicker. The
prediction of arrival times is known by everybody from the navigation sys- tems built in cars but tracking systems nowadays can do a lot more, they can provide real-time data for simulations and for example in case of heavy equipment calculating future productivity and by that helping decision mak- ing is possible. The customer service application has many sides too. In case of public transportation, there are well-established networks to provide in- formation about vehicles to the passengers. The possibility of tracking of the ordered or posted products during its way is also frequent at transport com- panies. But today it is far more than that, in Singapore limousines of a hotel are installed by vehicle tracking systems to make sure the welcome of their VIP guest by the moment of reaching the hotel [159]. The goal of this Chap- ter is to introduce the applied tracking systems especially by technological viewpoint in the case of two industries where it is frequently used, namely in transportation and construction.
3.1 About VTS and AVL in General
A VTS consists a software which gathers and uses the data of more than one automatic vehicle location (AVL) system installed in different vehicles, so it allows the possibility of monitoring multiple vehicles at the same time.
An automatic vehicle location system has four parts:
• Hardware ensuring the locating possibility
• Communication package which connects the individual units to the central server
• Computer display system
• Advanced components
The first three parts are indispensable, the fourth one is conditional de- pending of the use of the AVL (e.g.: engine monitoring elements).
Categorizing the AVLs by its locating hardware we can distinguish four basic types (Table3.1):
1. GPS-based: The Global Positioning System is a satellite navigation sys- tem which uses 24 satellites at about 20.000 km orbiting altitude. It was developed for military purposes but the civil utilization of it has been working for decades. (There is an alternative satellite system which is the GLONASS - Global Navigation Satellite System -.)
3.1. About VTS and AVL in General 29 2. Sign-post technology: The common idea of these technologies is the ap- plication of some kind of short range communication channels between the vehicle and the so called sign-post. This method tracks the vehicle when it is passing a given point.
3. Ground based radio: Ground based radio (GBR) technology utilizes one or more fixed land-based antennas to locate vehicles by radio tri- angulation. One example for its application is the Loran-C radio navi- gation system which uses low frequency radio signals.
4. Dead-reckoning (DR): The essential of the dead-reckoning is the prin- cipal if the starting position is known, the velocity (speed and orienta- tion) of the vehicle is continuously being recorded, so its position can be calculated. For this purposes different speedometers and compasses are installed in the machines [238].
An example for an approach which utilizes other tracking technologies than the above four basis ones is published by ([67]). This method was de- veloped for the specific purpose of law enforcement in urban environments and uses fixed cameras and unmanned aerial vehicles (UAV) to locate a ve- hicle and predict its motion.
TABLE3.1: Basic AVL types
Accuracy [m] Advantages Disadvantages
GPS-based <30 Flexible
Very accurate (with addition)
Signal interference Satellites are government properties
Sign-post 1-20 Accurate
Small interference
Fixed Maintenance
need
GBR 30-2000 Flexible Varying
accuracy Dead-
reckoning
<75 On vehicle Accurate on short distances
Measuring on each vehicle Inaccurate on large distances
Inspecting the AVL-s from the side of communication component we can classify them by technology and by data transferring time point of view.
Communication technologies:
• radio (analog/digital)
• cellular (analog/digital)
• satellite Data transfer:
• online (real-time)
• off-line
• quasi online
The online systems provide continuous data flow into the direction of server machine, while in the case of off-line systems, the recorded informa- tion is stored in a memory card and the reading of it happens when the ve- hicles arrives to the center. The drawback of the online systems is the fix communication cost which can be eliminated by using an off-line system;
however, it has much more human resource need. Beside that there are ap- plications which require real-time or near real-time data. An off-line so called quasi online system is implemented in [10] with automatized card reading at the garage (or even in stations). The quasi online feature means that the sys- tem can utilize the free wifi-spots for data transmission as well.
Once the vehicle is located visualizing the information on a map is the task of a computer display system. Digital maps can perform street geocod- ing, providing best route and guidance too. Nowadays Geographical Infor- mation Systems (GIS) can be used as computer display systems which results even more useful information, [84] based on this vehicle routing problems can be solved too [185].
In order to store the position data one can choose from the following op- tions:
• Direct position data storage: relational database (x,y coordinate, speed, moving direction, time, etc.), consumes very large space
• Motion functions description method: constructs a motion function which can help prediction (in urban areas it is complicated)
3.1. About VTS and AVL in General 31
• Spatial-temporal database: Store data in a space-time cube, in the field of spatial-temporal GIS, it is widely spread [221]
Most of the AVLs uses cellular communication technology with one of the most frequent applied location mechanisms which can be stand-alone GPS ([180]) or some hybrid of the basic location technologies. The mathematical basis of the GPS technology is the so called triangulation but in reality it is rather a complicated calculation process which considers the errors caused by the atmosphere, clocks, etc [217]. To improve the accuracy of GPS to a specific level different techniques are available for different applications:
• DGPS (Differential GPS): network of reference stations helps the posi- tioning up to 10 cm (requires DGPS station closer than 1000 km to the receiver)
• AGPS (Assisted GPS): uses auxiliary software in order to know the ex- act positions of the satellites, requires active data connection
• RTK (Real Time Kinematic): it provides cm-level precision by measur- ing the phase of the carrier wave of the signal
• e-Dif (extended Differential): useful at regions where differential cor- rections are not available ensuring 1 m accuracy
• other augmentation systems ([1])
There are algorithms also for successful data analysis, combining GPS sig- nal with external aids including Wiener filter, Kalman filter, neural networks, etc. Kalman filter which is an improvement of the Wiener one has the ad- vantage of requiring only the current measurement and state compared to Wiener filter. However, the Wiener filter proved to be the most precise one according to [69] comparison and using the fact that it is linear, the authors were capable of further improving its precision by using parallel architec- ture. Beside the accuracy another significant topic of nowadays researches is reducing the cost of GPS systems. [94] suggested the following for low cost Intelligent Vehicle Tracking Systems (IVTS) in urban environment:
• minimizing the calculations in vehicle,
• minimizing data transfer between in vehicle unit and the basis,
• under poor satellite visibility adopting alternative approaches.
3.2. Vehicle Tracking in Transportation 33 on two earlier ones which are the Nextbus, (it is a commercial passenger in- formation system applied in several cities in the USA) and the Mybus. The so called Mybus is a scientific project of University of Washington. This ar- rival time prediction system applies Kalman filter and divides the complete route into such small segments on which the travel speed can be considered constant. The Department of Transport Technology and Economics of the Bu- dapest University of Technology and Economics is also working on a project called BusEye. This passenger information system aims to improve the pub- lic transportation [28].
In public transportation in some cases of rail transport odometer and in- ertial sensors are sufficient to locate a train. But if we talk about public trans- port in urban environment it cannot be considered as 1D despite the fact of known path (but it can help the positioning). In these cases, GPS-based technology is applied generally but it has a remarkable drawback namely its price. [132] compared GPS technology to two sign-post technology, the first one is the so called vehicle card technique and the second one is Radio Fre- quency Identification (RFID), and RFID proved to be the best (Fig.3.1) in the investigation of the mentioned publication. The vehicle card technique is a method where a card containing vehicle data is required to carry to tracking points to transfer data, its disadvantage is the high scope of human interven- tion. The realization of the proposed method would contain RFID transmit- ters on the buses, receivers at the bus stops and Global System for Mobile Communication (GSM) connection to the server.
The recent developments of autonomous driving draws a lot from robotics, so new type of tracking methods are appearing on the roads. Simultaneous localization and mapping (SLAM) extends the odometry problem with map generation and uses loop closures to make the measurements more accurate.
SLAM can be based on any sensor signal about the environment which en- ables the registration of consecutive measurements and so the transformation matrix of rigid movement of mobile robots can be determined. Most frequent research topics about this field are camera based (visual) [138] and depth sensor based, e.g. Asus Xtion and LIDAR [206] SLAMs. Another emerging area is the Visible Light Communication (VLC) based localization, the main reason for that, there is still no appropriate alternative for GPS in case of in- door environment. [152] proposed a system where modulated LED lights are used as beacons to aid indoor localization. Further fresh indoor technologies are ’magnetic fingerprint’ based localization [30] - the position calculation is based on characteristics of magnetic fields - and ultrasound based one [172].
The application of VTSs during construction processes is relatively new, how- ever the elements of these systems have been present long time ago at con- struction sites. RTK GPS is applied in civil engineering as a high precision 3D measurement system offering good alternative instead of laser level systems.
Another example of the first uses of GPS in construction automation is the piling rig positioning in [184].
The recently applied technologies of Real Time Locating Systems (RTLS) at construction sites:
• GPS
• RFID: this technology is also used to allocate tools at the construction job site in order to get information about their availability [72] and for auditing the wearing of personal protective equipment (PPE) ([101])
• Ultra Wideband technology (UWB): the technology enables the trans- mission of large amount data using large spectrum of frequency and in resource location tracking in construction sites it has been already replacing the earlier used angle of arrival (AoA), time of arrival (ToA),