Precise Vehicle Positioning for Indoor Navigation via OpenXC

Yusuf Turk, Baturay Ozcan and Sezer Gören

Department of Computer Engineering, Yeditepe University, Kayışdağı, 34755, Istanbul, Turkey

Keywords: OpenXC, Indoor Positioning, Connected Car, Indoor LBS, Intelligent Parking Systems.

Abstract: We propose a method for vehicle positioning for indoor locations like parking garages. Our method does not

require other external positioning systems such as GPS. Instead, we determine the vehicle position from the

vehicle data obtained from an OpenXC dongle attached to the OBD-II interface of the vehicle. An

accompanying smartphone application which connects with the dongle via Bluetooth is developed. This

application calculates the position of the car and applies the algorithms proposed in this paper to the vehicle

data received from the interface. The route of the vehicle is then constructed and displayed on the smartphone

screen. As a future work, an assistant application will be developed that guides the driver to the spot where

the car was parked before.

1 INTRODUCTION

Finding parking spaces and locating the parked

vehicles in huge parking lots with multiple levels

such as large shopping malls, airports, and office

buildings is a common problem. Although traditional

solutions such as coloring and numbering the areas

worked until now, drivers still have to locate the car

without any assistance. Also, assisting the driver with

a single route to the location of the parked car is a

more time efficient solution.

Navigation systems in vehicles are becoming a

part of our daily life. Whether it is offered as an

embedded system included by the car manufacturer

or an external system using a smartphone, navigation

systems are more accessible than ever. Most of the

navigation systems use GPS which relies on

positioning values retrieved from various satellites.

This paper focuses on offering a route assistance

solution for the cases where GPS is not available such

as an indoor parking garage multiple levels below the

ground. Our solution is based on OpenXC platform

(OpenXC Platform, 2011). OpenXC vehicle interface

is composed of a microcontroller with two external

connections – one to the CAN bus via the OBD-II

port, and one to the host device via USB or Bluetooth.

It passively listens for a subset of CAN messages,

performs required unit conversion or factoring and

outputs a generic version to the USB interface. Many

vehicle output parameters such as vehicle speed,

steering wheel angle, door status, wind shield status,

etc. are available and sent through Bluetooth and

received by a custom smartphone application

developed on iOS or Android operating system.

OpenXC API enables the development of custom

applications using the information retrieved from the

vehicle. The adapter and the smartphone

communicate through Bluetooth. In this paper, we

developed a smartphone application which uses the

real-time values retrieved from OpenXC interface in

order to draw and save the route of the vehicle on the

indoor map of the parking garage and offers precise

vehicle positioning for indoor navigation. In the

future, our smartphone application can be updated to

guide the driver to the already parked car using the

saved route.

2 RELATED WORK

OpenXC is commonly used in vehicle systems along

with complementing tools and software. Android

application developed using OpenXC API can

retrieve real-time values from the vehicle to compare

them with threshold values (Suresh and Nirmalrani,

2014). Similarly, vehicle condition and sensor values

can be integrated with a network such as WiMAX in

order to create a networked Vehicle Sensing and

Control (VSC) system (Wang et al., 2016). Aside

from using OpenXC, vehicle data can also be

obtained by using in-vehicle eDSMS along with an

Android application (Akiyama et al., 2015).

440

Turk, Y., Ozcan, B. and Gören, S.

Precise Vehicle Positioning for Indoor Navigation via OpenXC.

DOI: 10.5220/0006770104400445

In Proceedings of the 4th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2018), pages 440-445

ISBN: 978-989-758-293-6

Regarding the indoor parking problem, many

solutions have been proposed before. While some of

them uses built-in car sensors, others use external

devices to determine the position of the vehicle.

Wagner et al. (2010) proposes a method using Dead

Reckoning (DR) and commercially available maps

from companies or from building plans. The method

utilizes a topological map matching algorithm that

combines geometrical information and the topology

of the road network. Bojja et al. (2013) takes

advantage of low cost gyroscope and odometer

sensor. Sensor data is combined with the 3D map of

the building to obtain the indoor position. Razak et al.

(2015) tried to navigate the driver to an appropriate

parking position using QR codes placed in the

parking space. Android application reads the QR code

and navigates the driver to a parking spot using the

map stored in the database. Vehicle movement is

calculated using a step counter. Kumar et al. (2016)

mounted cameras throughout an indoor parking lot to

identify vehicle movement. The data was used to

calculate vehicle position by using convolutional

neural network with Deep Learning.

Indoor positioning and parking services also use

smartphone sensors to determine the position of the

vehicle. Some studies combined smartphone sensor

data with RSSI data of WLAN and the Bluetooth

signal data. Liu et al. (2012) proposed a cost and

energy efficient solution. Wireless Local Area

Network (WLAN) signals and the built-in

smartphone sensor data combined to find the precise

location of the vehicle inside an indoor parking space.

In addition to that, another study by Liu et al. (2012)

also used WLAN signals and smartphone sensors. An

indoor positioning engine is developed for the

positioning. Motion information is measured using

smartphone sensors. The solution can be used for

real-time navigation and location tracking. Similarly,

Bluetooth signals strengths can also be used for

location fingerprinting which is combined with

relative positioning in a study by Wilfinger et al.

(2016). Bluetooth Low Energy (BLE) is offered as a

cost-effective solution. Khare et al. (2017) stated that

the BLE and IEEE 802.15.4a compliant Ultra-

wideband (UWB) sensor data can be processed in

real-time to develop an indoor localization system.

Unfortunately, UWB sensors are not in current

handheld devices yet.

Research by Li et al. (2016) is focused on car

searching system that will help users to find their cars

in a large parking lot. QR codes are distributed over

the parking lot and a smartphone application

calculates the path to the vehicle using the shortest

path algorithm.

Internet of Things (IoT) based applications and

middleware systems are also used for parking

solutions for both current use cases and smart cities.

Ji et al. (2014) stated that car parking services must

be intelligent for IoT based smart cities. They

proposed a cloud based system offering users a car

parking service experience. In an another IoT

application by Bagula et al. (2015), RFID readers are

distributed in a parking lot to help billing and provide

smart parking. Similarly, Caballero-Gil et al. (2016)

used Near Field Communication (NFC) tags for their

indoor positioning system.

iBeacon is another technology which offers a low

energy solution compared to GPS and WLAN. Zou et

al. (2016) used iBeacons to provide positioning and

navigation for vehicles using both indoor and outdoor

spaces.

Indoor location positioning and navigation is also

used for autonomous car parking systems. A study by

Correa et al. (2017) proposes a cooperative vehicular

positioning network using the Vehicular Sensor

Networks (VSN).

3 PROPOSED METHOD

OpenXC supports various number of measurements

of a car such as engine speed, steering wheel angle,

fuel consumption, accelerator pedal position, ignition

status, etc. The data retrieved using the OpenXC API

as JSON formatted messages in a class called Vehicle

Messages. In our design, we make use of three of the

measurement data received through OpenXC vehicle

interface: i. steering wheel angle, ii. odometer, and iii.

ignition status. Steering wheel angle provides the

information about the angle of the steering wheel in

degrees in a range from -600 to +600. When the

steering wheel is turned to the right, positive values

are read in the message and negative when to left.

Odometer gives the data of the distance that the

vehicle travels as a unit of kilometer between the

values 0 to 16777214 with about 0.2-meter resolution.

Ignition status returns as a Boolean value indicating

one of the 4 states of the ignition: off, accessory, run,

and start.

In our implementation, we have three functions: i.

CheckTurn (Figure 1), ii. UpdateTurn (Figure 2), iii.

UpdateDirection (Figure 3). CheckTurn function

checks whether the car has turned or not. If so, the

route is updated. At the beginning, we consider the

ignition status of the vehicle. If it is off and the car is

in the parking area, we know that the driver has

parked his/her car in a parking spot. Therefore, we

update the route and return.

Precise Vehicle Positioning for Indoor Navigation via OpenXC

441

Figure 1: Algorithm for CheckTurn Function.

In order to detect a turn, we set a threshold value

for the steering wheel angle. If the current angle of

the steering wheel exceeds the threshold we assume

turn is started or continues when it has already started.

Exceeding the threshold means that the angle can be

over the positive threshold or below the negative

threshold. When this is the case, we check the

startRouting value which indicates if the car has

entered the parking area or not. If false, just return.

Otherwise, if the vehicle has already started to turn,

we set turnContinue value as true. If not, that means

rotation is just started, then we set turnStart value as

true. If the threshold value is not exceeded but the car

was in turning state, we predict that the rotation is

over and update the map. To do this, we need the

other two functions: updating the coordinates with

turn points and updating the direction after the turn.

The threshold value can be adjusted to detect the

different types of curves. For example, a U-turn is

detected when there are multiple and complete

steering wheel rotations.

UpdateCoord (Figure 2) is a function which we

save the coordinates of the turning points. It takes the

parameter distance (value in the unit of meters

showing the distance between the previous turn point

and the current one) and updates x and y coordinates

of the current turn point. For example, if the direction

Figure 2: Algorithm for UpdateCoord Function.

of the vehicle was north before the rotation, x

coordinate does not change but y coordinate is

decreased.

In UpdateDirection function (Figure 3), we set the

direction of the vehicle after the turn which is a

critical issue. Our solution runs in a way that we

check two values together: steering wheel angle and

the direction of the vehicle before the turn. First, we

find to which direction the vehicle is rotated. If the

value of the angle is positive, we predict that it is

rotated to the right. Otherwise, to the left. Then, we

take the previous direction of the vehicle under

VEHITS 2018 - 4th International Conference on Vehicle Technology and Intelligent Transport Systems

442

consideration. For example, if the car is rotated to the

left and the previous direction is east, it is clear that

the driver was going to the east and then he/she turned

to the left, so the new direction of the car is now north.

4 EVALUATIONS

We evaluated the system using an OpenXC adapter

attached to a car that supports the adapter and a

smartphone with Android operating system. The car

was driven to a small parking garage with the parking

plan given in Figure 4. In the figure, the color white

represents the roads and the grey rectangles represent

the parking spaces. The entrance of the parking lot is

the road at the southern part of the map. The map is

drawn proportionally.

The car was driven inside the indoor parking lot

until an empty parking spot is found. The OpenXC

adapter used in tests does not support GPS and all the

location services including the GPS was disabled on

the smartphone. The smartphone application was run

when the car entered the parking lot. Although a static

map is used in our system, it can be replaced according

to the last known location retrieved by the smartphone.

Figure 5 shows the car movements captured by the

application in real time. A new route is drawn on top

of the map whenever a new turn is detected. Route is

smoothened by the application in order to create an

easy to understand style for the map. Parking sign at

the bottom indicates the entrance of the indoor

parking lot. The car icon indicates the location of the

car and it is updated in real time. When the car is

parked, the car icon is moved inside the correct

parking spot. Smoothing the route created according

to the data from the OpenXC allowed the application

Figure 3: Algorithm for UpdateDirection Function.

to draw straight lines from corner to corner. Since the

most parking spaces in office buildings, shopping

malls, and airports consists of proportionally

designed rectangular parking spots, it will be

sufficient for most of the use cases. While we were

doing the tests, we always observed that the vehicle

and parking lot sizes are detected with an error of less

than a half meter. In addition, we also observed that

the route was successfully drawn without crossing the

area of the parking space.

Figure 4: Blueprint of the car park.

The steering wheel angle range from -600 to 600

where positive values indicate that wheel is turning to

the right and negative values indicate that the wheel

is turning to the left. If the driver turns left, it does not

necessarily mean that the steering wheel angle values

will always be negative. This is because the steering

action can consist of steering to the right for a certain

angle, followed by a large angle steering to the left.

Among all the in-car tests, our algorithm calculated

the steering angles and turn direction correctly. The

application saves the location of the parking spot

when the ignition is switched off.

5 CONCLUSIONS

In this paper, we proposed a vehicle positioning

method for the indoor navigation without the help of

external positioning systems, but with OpenXC

vehicle data. The indoor route of the vehicle is built

and saved by our smartphone application. As a future

work, driver assistance to the last parked location of

the vehicle will be added as a new feature to our

application. In addition, detection of the story level in

multi-story parking garages is another future work.

We expect that smart parking solutions and assistants

like our approach will be more common in the

following years.

Precise Vehicle Positioning for Indoor Navigation via OpenXC

443

Figure 5: Car Movements in Real Time.

REFERENCES

Openxcplatform.com, 2011. OpenXC Platform. [online]

Available at: www.openxcplatform.com/ [Accessed 3

November 2017].

V. Suresh and V. Nirmalrani. (2014). "Android based

vehicle diagnostics and early fault estimation system".

In: International Conference on Computation of Power,

Energy, Information and Communication (ICCPEIC).

Chennai, 2014, pp. 417-421.

Y. Wang et al. (2016). "VLnsight: enabling open

innovation in networked vehicle sensing and control".

In IEEE Network, vol. 30, no. 4, pp. 34-44, July-August

2016.

S. Akiyama, Y. Nakamoto, A. Yamaguchi, K. Sato and H.

Takada. (2015). “Vehicle Embedded Data Stream

Processing Platform for Android Devices”. In:

International Journal of Advanced Computer Science

and Applications (IJACSA), 6(2).

J. Wagner, C. Isert, A. Purschwitz and A. Kistner. (2010)

"Improved vehicle positioning for indoor navigation in

parking garages through commercially available maps".

In: 2010 International Conference on Indoor

Positioning and Indoor Navigation, Zurich, pp. 1-8.

J. Bojja, M. Kirkko-Jaakkola, J. Collin and J. Takala.

(2013). "Indoor 3D navigation and positioning of

vehicles in multi-story parking garages". In: 2013 IEEE

International Conference on Acoustics, Speech and

Signal Processing, Vancouver, BC, 2013, pp. 2548-

2552.

S. F. A. Razak, C. L. Liew, C. P. Lee and K. M. Lim.

(2015). "Interactive android-based indoor parking lot

vehicle locator using QR-code". In: 2015 IEEE Student

Conference on Research and Development (SCOReD),

Kuala Lumpur, 2015, pp. 261-265.

A. K. T. R. Kumar, B. Schäufele, D. Becker, O. Sawade and

I. Radusch. (2016). "Indoor localization of vehicles

using Deep Learning". In: 2016 IEEE 17th

International Symposium on A World of Wireless,

Mobile and Multimedia Networks (WoWMoM),

Coimbra, 2016, pp. 1-6.

J. Liu, R. Chen, Y. Chen, L. Pei, L. Chen (2012). “iParking:

An Intelligent Indoor Location-Based Smartphone

Parking Service”. In: Sensors 2012, 12, 14612-14629.

J. Liu, R. Chen, L. Pei, R. Guinness, H.A. Kuusniemi,

(2012). “Hybrid Smartphone Indoor Positioning

Solution for Mobile LBS”. In: Sensors 2012, 12, 17208-

17233.

R. Wilfinger, T. Moder, M. Wieser and B. Grosswindhager.

(2016). "Indoor position determination using location

fingerprinting and vehicle sensor data". In: 2016

European Navigation Conference (ENC), Helsinki,

2016, pp. 1-9.

S. P. Khare, J. Sallai, A. Dubey and A. Gokhale. (2017).

"Short Paper: Towards Low-Cost Indoor Localization

Using Edge Computing Resources" 2017 IEEE 20th

VEHITS 2018 - 4th International Conference on Vehicle Technology and Intelligent Transport Systems

444

International Symposium on Real-Time Distributed

Computing (ISORC), Toronto, ON, 2017, pp. 28-31.

J. Li, Y. An, R. Fei and H. Wang. (2016). "Smartphone

based car-searching system for large parking lot" 2016

IEEE 11th Conference on Industrial Electronics and

Applications (ICIEA), Hefei, 2016, pp. 1994-1998.

Z. Ji, I. Ganchev, M. O'Droma, L. Zhao, X.A. Zhang.

(2014). “Cloud-Based Car Parking Middleware for

IoT-Based Smart Cities: Design and Implementation”.

Sensors 2014, 14, 22372-22393.

A. Bagula, A, L. Castelli, M. Zennaro. (2015). “On the

Design of Smart Parking Networks in the Smart Cities:

An Optimal Sensor Placement Model”. Sensors 2015,

15, 15443-15467.

C. Caballero-Gil, P. Caballero-Gil, J. Molina-Gil. (2016).

“Cellular Automata-Based Application for Driver

Assistance in Indoor Parking Areas”. Sensors 2016, 16,

1921.

H. Zou, H. Jiang, Y. Luo, J. Zhu, X. Lu, L. Xie,.

(2016). “BlueDetect: An iBeacon-Enabled Scheme

for Accurate and Energy-Efficient Indoor-Outdoor

Detection and Seamless Location-Based Service”.

Sensors 2016, 16, 268.

A. Correa, G. Boquet, A. Morell, J. Lopez Vicario. (2017).

“Autonomous Car Parking System through a

Cooperative Vehicular Positioning Network”. Sensors

2017, 17, 848.

Precise Vehicle Positioning for Indoor Navigation via OpenXC

445