Design, Development and Testing of Automatic

Pothole Detection and Alert System

Dhwani Desai1, Abhishek Soni2, Dhruv Panchal3, Sachin Gajjar*
Institute of Technology, Nirma University, Ahmedabad, Gujarat, India
mail:[email protected],[email protected], [email protected], *corresponding author: [email protected]

Abstract—Every year several road accidents take place solely II. HARDWARE DESIGN
credited to the potholes. In this paper, an Automatic Pothole
Detection and Alert System is developed using ultrasonic sensor, A. Microcontroller
accelerometer, stereo camera and Global Positioning System Raspberry Pi (RPi) 3B+, the single board computer with ARM
(GPS) integrated with Raspberry Pi. The system identifies a Cortex-A53, on board memory of 1GB SRAM, Wi-Fi,
pothole, alerts the rider and generates location database of existing Bluetooth and Ethernet connectivity is used as microcontroller
potholes. This in turn can create cautiousness in the riders, for the system [9].
decrease number of accidents and maintenance costs of vehicles.
The experimental results state that the system is able to identify B. Accelerometer
the potholes with 90% accuracy. The cost of the system is around MPU 6050 consisting of six axes gyroscope and accelerometer
8000 INR and it reduces average vehicle maintenance cost by 35- is used [10] to measure depth of the pothole. It is a micro-
50 %. The time taken to send GPS location to the database and electromechanical system motion tracking device with I2C
send alert to the rider is 4 sec. A strong, city-wise database is
interface.
generated which can further chisel the system for the entire nation.
C. Ultrasonic sensor
Keywords— Automatic Pothole Detection, Hardware Design,
Raspberry Pi, Stereo Camera, Ultrasonic Sensor, Accelerometer. Ultrasonic sensor HC-SR04 is used to measure depth of pothole
and the height of speed breakers along with their distance from
I. INTRODUCTION the vehicle [11]. It consists of ultrasonic waves transmitter,
According to the recent reports of Ministry of Road Transport receiver and control circuit. Fig. 1 shows the working of
and Highway, Government of India there were approximately ultrasonic sensor with the threshold fixed to a distance of 400
9,300 deaths in 2015, 2016 and 2017 in road accidents due to cms. When the distance detected is less than threshold, it
indicates a speed breaker and when it is above, it indicates a
potholes and 25,000 people were injured [1]. For safety of
pothole. Using IO trigger a high level signal is transmitted for
vehicle riders, a number of researches to develop an automatic 10 µs. The module sends eight 40 kHz signals and detects pulse
pothole detection and alert system is carried out. Work in [2] signal back. If the signal is back, time of high output IO
and [3] proposes a system with an ultrasonic sensor to detect duration is the time from sending ultrasonic to returning back.
presence of pothole and speed breakers on the road. Work in [4] The distance is calculated using equation (1).
and [5] uses accelerometer of mobile phones to detect
High Level Time x Velocity of Sound
vibrations in the vehicle to identify the potholes. Work in [6] Test Distance = (1)
2
uses LiDAR and that in [7] uses LASER to identify a pothole.
However, the cost of implementation of both the approaches is
nearly 50,000 INR. Work in [8], proposes a system consisting
of arduino, ultrasonic sensor, GPS and accelerometer to identify
potholes. All the above mentioned approaches do not provide
multiple needs of the system like: detection of the pothole;
alerting the driver; identifying the height of speed breaker,
depth of pothole and their location; creating a database of
location of potholes for alerting future riders and government
bodies to take repair actions. To this end, this paper introduces
a system that encompasses all of the above needs of the system.
Above all, the system is cost efficient and can be easily fitted in
the vehicles. The rest of the paper is organized as follows. The
Fig 1. Working of Ultrasonic Sensor
hardware design and working of the system is discussed in
Section II and III, respectively. Experiments performed on the D. Stereo Camera
system and the related results are presented in Section IV.
Sony IMX219 ISX, a stereo vision camera with pixel rate of
Finally the paper is concluded in Section V.
280 Mega Pixels/secs in all pixel mode is used [12]. To analyze
size of pothole and height of speed breaker, Intel Neural
compute Stick, a vision processing unit is used along with the
camera [13]. For detecting the depth of the pothole, adaptive
breakpoint detection is used [14]. The camera is mounted in a

978-1-7281-2327-1/19/$31.00 ©2019 IEEE

Authorized licensed use limited to: University of Exeter. Downloaded on May 08,2020 at 10:08:25 UTC from IEEE Xplore. Restrictions apply.
way that it can analyze the depth and size of a pothole while a
vehicle is crossing over it.
E. Global Positioning System
GPS u-blox neo 6M is used to identify the location (latitudes
and longitudes) of pothole or speed breaker [15]. It has a built
in 25 x 25 x 4 mm ceramic antenna for searching satellite and a
data backup battery to store location information in case of
accidental shut down with a ROM7.03 storage.
III. WORKING OF THE MODULE
Fig. 2 shows the working of the system. Accelerometer and
ultrasonic sensor identify the pothole depth and distance. The
size of the pothole is calculated by using the camera and the
neural compute stick and its location by GPS. Next Simple Mail
Transfer Protocol (SMTP) is used to send E-mail notification to
Fig 3. Alert notification of Pothole and GPS Location of Pothole on Google
the rider. The ultrasonic sensor calculates distance Maps
continuously. The threshold distance set is 400 cms and the
sensor keeps a check whether the distance is fixed or not.
Whenever there is any change in the distance measured with the
received ultrasonic waves, it alerts the rider, that there is a
pothole or a speed breaker affecting the ride. The accelerometer
works when the vehicle is crossing over the pothole. It
continuously keeps a check over the altitude of the suspension
and when there is a pothole it calculates its depth. Based on the
depth of the pothole it is decided whether the pothole is major
or minor in terms of depth. The stereo camera is mounted at the
front panel of the vehicle and the size of the pothole is analyzed
using neural computer stick. Whenever a pothole is detected,
GPS collects longitudes and latitudes of the location and stores
it. SMTP is used to send E-mail alert to the registered e-mail
address of the rider with location and depth of the pothole as
shown in the Fig. 4. The data is also uploaded on the server
which can further be utilized by the government to initiate the
repair work and alert the other riders through google maps as
shown in the Fig 3. Fig. 4. System mounted on a two wheeler

at 5.5 minutes shows the occurrence of a pothole. As shown in

Alert Notification to
Riders the Fig. 6, the threshold for the pothole is considered to be 5
cms [1]. At 5.5 minutes, the distance measured by ultrasonic
Governing
sensor is greater than the threshold indicating a pothole
Server Database/
Cloud Bodies detection. Since there is no pothole detected before 5.5 minutes,
the depth of the pothole is 0. But at 5.5 minutes the depth of the
pothole is measured as 28.3 cms. Thus the ultrasonic sensor not
Raspberry Pi 3b+ GPS Receiver only detects the pothole but also measures its depth. Fig. 7
shows the acceleration in z direction. A constant acceleration in
Ultrasonic Sensor Accelerometer z direction shows that vehicle is moving on a smooth road i.e.
Stereo without any pothole. At 6th minute, variation is observed in
Vision
Camera
acceleration denoting that the vehicle is crossing a pothole. This
detection of pothole is not useful to the current rider but can be
used for other riders, to alert them in advance regarding the
Fig 2. General Flow Chart of the system
pothole. Whenever the ultrasonic sensor and accelerometer
IV. EXPERIMENTAL RESULTS detects a pothole, it triggers the camera which clicks a picture
of the pothole. The picture along with pothole depth and
Fig. 4 shows the system mounted on a two-wheeler for location is sent to a server. Fig. 8 shows power consumption of
experimentation. As seen in the Table I, the system gives individual components of the system. Several experiments on
accurate details of potholes and speed breakers along with their the roads with vehicle speed ranging from 10 to 50 kmph show
depth, height, longitudes and latitudes. Fig. 5 shows the that the system provides accuracy of 90%. Avoiding potholes
distances between the vehicle and the road at different time and speed breakers leads to monetary savings of 35-50 % on
instances taken by the ultrasonic sensor. In the figure, the peak vehicle maintenance.

Authorized licensed use limited to: University of Exeter. Downloaded on May 08,2020 at 10:08:25 UTC from IEEE Xplore. Restrictions apply.
 Sr. Obstacle Height/Depth Latitude Longitude Power Consumptions (mW)

No. Type (cms) (degrees) (degrees)

1 Pothole 19.35 23.9563 72.5544 GPS 100

2 Speed Breaker 3.1 23.9406 72.5661

3 Speed Breaker 3.8 23.9421 72.5668 16x2 LCD 25

4 Pothole 13.2 23.9434 72.5669

5 Pothole 8.7 23.9411 72.5654 Accelerometer 16

6 Pothole 6.3 23.9423 72.5658

7 Speed Breaker 2.3 23.9547 72.5755 Ultrasonic Sensor 75

8 Pothole 15.8 23.9575 72.5769

9 Speed Breaker 3.1 23.9417 72.5659 Raspberry Pi 3b+ 2000

10 Pothole 18.2 23.9567 72.5760

0 250 500 750 1000 1250 1500 1750 2000 2250
Table I. Data collected from ultrasonic sensor and GPS module
Series 1
Distance Measurement by Ultrasonic Sensor
Fig. 8 Power consumption of the components of the system
m et a- char t .com
35

30
The nation-wide implementations require some modifications
such as, accounting weather conditions and road restrictions.
25
The database of every city along with road conditions can be
Distance (cms)

20 created with the help of government portals. Thus, with

15
database creation, the module can be developed as an additional
and necessary safety feature in modern vehicles.
10

5 REFERENCES
0 [1] Government of India Transport Research Wing Report, "Road Accidents
0 1 2 3 4 5 6 7
Time (minutes) in India", Ministry of Road Transport and Highways [Online], 10 June
2019. Available: https://www.morth.nic.in/about-road-safety.
Fig 5. Detection of Pothole by using ultrasonic sensor [2] R. Madli, S. Hebbal, P. Pattar and V. Golla, "Automatic Detection and
Pothole Detection and Depth Measurement by Ultrasonic Sensor
Notification of Potholes and Humps on Roads to Aid Drivers," in IEEE
30
Sensors Journal, vol. 15, no. 8, pp. 4313-4318, Aug. 2015.
25
[3] S. Srivastava, A. Sharma and H. Balot, "Analysis and Improvements on
Current Pothole Detection Techniques," 2018 International Conference
20 on Smart Computing and Electronic Enterprise (ICSCEE), Shah Alam,
Pothole Depth (cms)

Malaysia, 2018, pp. 1-4.

15 [4] V. Rishiwal and H. Khan, "Automatic pothole and speed breaker
detection using android system," 2016 39th International Convention on
10
Information and Communication Technology, Electronics and
5 Microelectronics (MIPRO), Opatija, Croatia, 2016, pp. 1270-1273.
[5] P. M. Harikrishnan and V. P. Gopi, "Vehicle Vibration Signal Processing
0 for Road Surface Monitoring," in IEEE Sensors Journal, vol. 17, no. 16,
pp. 5192-5197, 15 Aug.15, 2017.
-5
0 1 2 3 4 5 6 7 [6] B. Kang and S. Choi, "Pothole detection system using 2D LiDAR and
Time (minutes) camera," 2017 9th International Conference on Ubiquitous and Future
Fig 6. Pothole depth measurement by ultrasonic sensor Networks (ICUFN), Milan, Italy, 2017, pp. 744-746.
Acceleration in Z direction by Accelerometer
[7] K. K. Vupparaboina, R. R. Tamboli, P. M. Shenu and S. Jana, "Laser-
0.6
based detection and depth estimation of dry and water-filled potholes: A
geometric approach," 2015 Twenty First National Conference on
0.4
Communications (NCC), Mumbai, India, 2015, pp. 1-6.
[8] C. Chellaswamy, H. Famitha, T. Anusuya and S. B. Amirthavarshini, "IoT
Accelerations (m/s2 )

0.2
Based Humps and Pothole Detection on Roads and Information
Sharing," 2018 International Conference on Computation of Power,
0
Energy, Information and Communication (ICCPEIC), Chennai, India,
2018, pp. 084-090.
-0.2
[9] Raspberry Pi Trading Ltd., "Raspberry Pi 3 Model B+", Raspberry Pi
-0.4
Foundation [Online], 5 March 2019. Available: https://www.raspberrypi.
org/products/.
-0.6
[10] TDK Corporation, "MPU-6050 Six-Axis (Gyro + Accelerometer)
0 1 2 3 4 5 6 7 8 MEMS", MotionTracking Devices TDK Corporation [Online], 5 March
Time (minutes)
2019. Available: https://www.invensense.com/products/motion-tracking/
Fig 7. Variation in axis measured by accelerometer 6-axis/mpu-6050/.
[11] Elprocus, "HC-SR04 Reference", Ultrasonic Detection – Basics and
V. CONCLUSION Application Elprocus [Online], 5 March 2019. Available: https://www.
elprocus.com/ultrasonic-detection-basics-application/.
A system that provides a better road experience to the rider by [12] IMX219PQ Reference Manual, "CMOS Image Sensor for Mobile
identifying a pothole, alerting the rider and generating location Products", Sony Semiconductor Solutions Corporation [Online], 2 March
database of existing potholes is designed, developed and tested. 2019. Available: https://www.sony-semicon.co.jp/products_en/
[13] Intel Software Developer Zone. (2019, March 5th), Movidius Neural
The system can be fitted in any two or four-wheeler vehicle and Compute Stick, Intel Corporation [Online]. Available:
can prove to be a national guide to safety on roads. https://software.intel. com/en-us/movidius-ncs

Authorized licensed use limited to: University of Exeter. Downloaded on May 08,2020 at 10:08:25 UTC from IEEE Xplore. Restrictions apply.
[14] G.A. Borges, MJ. Aldon, "Line Extraction in 2D Range Images for
Mobile Robotics," Journal of Intelligent and Robotic Systems, vol. 40
Issue 3, pp. 267-297, July 2004.
[15] u-blox. (2018, Nov. 11). NEO-6 series Versatile u-blox 6 GPS modules
[Online]. Available: https://www.u-blox.com/en/product/neo-6-series

Authorized licensed use limited to: University of Exeter. Downloaded on May 08,2020 at 10:08:25 UTC from IEEE Xplore. Restrictions apply.