FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
The Patents Rules, 2003
COMPLETE SPECIFICATION
(See section 10 and rule 13)
1. TITLE OF THE INVENTION
Neural Network Based Interlinked Motorcycle Ecosystem For Smart Navigation.
2. APPLICANT(S)
(a) NAME
(b) NATIONALITY
(c) ADDRESS

Mr. Rohan R. Mahajan Indian 012A- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector-7, Phase -1, Nerul, Navi Mumbai
Mr. Viraj S. Rane Indian 012A- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector- 7, Phase -1, Nerul, Navi Mumbai
Mr. Heramb S. Khanvilkar Indian 012A- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector- 7, Phase -1, Nerul, Navi Mumbai
Mr. Kartikey A. Kawadkar Indian 012A- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector- 7, Phase -1, Nerul, Navi Mumbai
Mr. Sachin R. Iyer Indian 012A- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector-7, Phase -1, Nerul, Navi Mumbai
Mrs. Shweta S/ Ashtekar Indian 205- Ramrao Adik Institute of Technology, Dr. D.Y. Patil Vidyanagar, Sector-7, Phase -1, Nerul, Navi Mumbai
3. PREAMBLE TO THE DESCRITION
COMPLETE
The following specification particularly describes the invention and the manner in which it is to be performed
4. DESCRIPTION
As attached on page 02
5. CLAIMS
As attached on page 07
6. DATE AND SIGNATURE
As attached on page 06
7. ABSTRACT OF THE INVENTION
As attached
Note:
*Repeat boxes in case of more than one entry.
*To be signed by the applicant(s)or by the authorized registered patent agent.
*Name of the applicant should be given in full, family name in the beginning.
•Complete address of the applicant should be given stating with postal index no. / code,
state and country.
*Strike out the column which is/are not applicable.

TITLE of Invention: Neural Network based Interlinked Motorcycle Ecosystem for Smart Navigation.
1. BACKGROUND OF THE INVENTION
Smart Navigations systems like assisted riding and autonomous riding fails in India due to the potholes and speed-breakers. Deaths and gruesome Injuries due to such deformities are quite high. Also, each and every deformity is different from each other. The dynamics of these deformities are quite random. This makes it even difficult to identify a given deformity from a pre-set threshold.
Motorcycles today come in various shapes and sizes. The proposed invention aims at a crowd sourced ecosystem environment where the motorcycle riders share the mutual ride data to each other. Solutions to identifying deformities can be using image recognition and processing, but such a system becomes expensive and not a practical approach for day-to¬day use. Indian environment being dusty and humid makes it an improper choice to use expensive and environment sensitive hardware onto a motorcycle.
The device is implemented onto a PCB which intentionally has 2 microcontrollers with Wi-Fi capabilities that process, store and upload the data onto the cloud server. The on-site neural network processing would require a powerful and more power consuming micro¬controller; hence the neural network resides onto the main Sever.
Smartphones today are the most powerful pocket computing devices. Also the availability of application development resources and tools makes the smartphone the best choice for visual aid in our context of smart navigation.
Existing technologies for identification of deformities using IMU sensors provides the solution to identify any one of the deformity like a pothole or a speed-breaker as the obtained acceleration versus time profile provides with just the presence of a deformity indicated by an impulse spike. So a different identification mechanism for differentiating amongst the pothole and speed-breaker must be developed. The proposed invention does the same.
The identification of the deformity happens if a rider goes through the pothole or over the speed-breaker. Then the data is processed and put onto the cloud server; from where the data is fetched and plotted onto the android application.
We also have a mechanism for detection of potholes if the rider doesn't go through it. For a set of 10 riders, if 7 riders change their lane which indicates a change in Y-axis of the motorcycle; in a range of 10 meters from the GPS location point where the first rider changes his lane then it indirectly indicates that there is a deformity. But even if one of the riders goes straight through then the presence of the deformity is confirmed.

2. DETAILED DESCRIPTION OF INVENTION
Block diagram of the proposed invention is shown in Figure 1. The data from the IMU sensor i.e. MPU6050 is given to the main On-board computer i.e. Raspberry Pi zero through ESP32. The ESP32 runs the regression algorithm which curve fits the data from the 3 IMU sensors.
This regressed data goes to the Raspberry Pi zero which runs the k-means clustering algorithm. Clustering is done to obtain relevant data windows and to get rid of unnecessary logged data. In order to avoid false logging of data due to swerving roads, the suspension travel of the motorcycle is calculated. A Pair of Ultrasonic sensors i.e. HC SR 04 is used to calculate the suspension travel. When the front suspension detects a travel more than a pre-set value then it indicates true data set logged from the IMU sensors.
The logged and k-means processed data is uploaded to the cloud server. This data is now fed to the neural network. The proposed invention uses a feed forward network with 3 input planes, 2 output planes. Now a speed-breaker or a pothole is identified. This data is matched with the GPS log file fetched from the android smartphone using timestamps for each of the data set and the GPS co-ordinate. This is now plotted onto the Map interface in the android application as shown in Figure 4.
When the rider rides with the smartphone mounted onto the motorcycle, he gets the notification visually and via speech if he is approaching a deformity. If repair for a deformity is done, then the new ride data from all the riders is cross referenced every 7 days to check for the same. If the new data and old data do not overlap in certain area, it indicates repair has been done at that GPS point.
2.1 PROPOSED METHODOLOGY AND WORKING
The proposed invention works by obtaining data in XZ plane and if triggered then in XY plane. The data from the XY plane system is cross referenced with the XZ plane data for that GPS co¬ordinate range of 10 meters. The system is inspired by the smart navigation systems of TESLA and BMW automobiles. The setup is a smart, low power consuming and compact device fitted under the seat of a motorcycle. All the data from the sensors needs to be processed depending on the data clusters. This is because in India we have different types of roads made of tarmac, paver blocks, gravel etc. So for each different material, the resonant frequency is different, so the noise added onto the IMU sensor's data is different. Hence, processing of the data is required. We use Regression for curve fitting the raw data and then for vector quantization we use k-means clustering.

2.2 Programming
Flow chart for program flow is shown in figure 3. The Android application is developed using android studio. It is built on top of Google Maps API, which also provides the point-to-point navigation and traffic condition updates. The android application receives data via Wi-Fi. The server sends data classified into 2 categories i.e. speed-breaker and pothole. The data is just GPS co-ordinate. Thus the app plots a red marker for the pothole and a blue marker for speed-breaker. The ultrasonic sensors give suspension travel in centimetre. This is the trigger point for true data set logging marker.
The Raspberry Pi zero connects to the cloud server via internet provided by any portable hotspot source.
2.3 Hardware Details
The circuit for the proposed invention is manufactured onto a printed circuit board. The PCB houses an ESP32 module, a Raspberry Pi zero, 3 connector ports for MPU6050s, 2 ports for Ultrasonic sensors and an activity indicating LED. Hardware and Software integration Block diagram is shown in Figure 2. The circuit takes power from a secondary battery bank. The secondary battery bank is charged from the main motorcycle power generator. All the IMU sensors operate on 3.3v. The Ultrasonic sensors operate on 5v.
The 3 IMU sensor nodes are mounted in front, centre and the rear of the motorcycle respectively. The Ultrasonic sensors are connected to a base plate and mounted under the T-fork in between the front suspensions. A buck boost converter is used in accordance to charge the secondary battery from the main battery system. The main server is a highly efficient computing machine with heavy processing capabilities and continuous internet connectivity.
3. SAMPLE RESULTS OBTAINED USING THE SETUP
The proposed system is able to provide speech and visual indication of deformity before the rider approaches it. For a new deformity, if the rider goes through then the activity gets recorded and then it is processed and uploaded onto the android application.
When road repairs take place and deformity like pothole is filled or a speed-breaker is flattened, data set of riders at that GPS location for a time period of 7 days is overlapped. If the old and new data doesn't overlap, then the repairs have been indicated onto the application by removing the deformity marker. Figure 6 and Figure 6 show the markers indicating pothole and speed-breaker respectively.
For an instance in a wide road, if the rider sees a pothole and avoids it by changing the lane, this is a change in Y axis of the motorcycle. For 7 out of 10 riders, if the similar Y axis change is

observed then it indirectly confirms a deformity. If at least one of the rider goes through straight then it is confirmed if the deformity is present or not by the change in Z axis of the motorcycle.
References:
[1] Ruicheng Zhong5 Guoliang Li, Kian-Lee Tan, Lizhu Zhou, Zhiguo
Gong, "G-Tree: An Efficient and Scalable Index for Spatial Search on Road Networks", Knowledge and Data Engineering IEEE Transactions on, vol. 27, pp. 2175-2189,2015, ISSN 1041-4347.
[2] Li Peng, Huang Xinhang, Wang Min, "A Hybrid Method for Dynamic Local Path Planning", Networks Security Wireless Communications and Trusted Computing 2009. NSWCTC '09. International Conference on, vol. 1, pp. 317-320,2009.
[3] M. Geetha, G. M. Kadhar Nawaz, "Fuzzy-ant based dynamic routing on large road networks", Pattern Recognition Informatics and Medical Engineering (PRIME) 2012 International Conference on, pp. 172-177, 2012.
[4] Aerodynamics of Road Vehicles (Premiere Series Books) 28 Feb 1998 by Wolf-Heinrich Hucho.
[5] Yazhe Wang, Baihua Zheng, "Hypergraph index: an index for contextaware nearest neighbor query on social networks", Social Network Analysis and Mining, vol. 3, pp. 813,2013, ISSN 1869-5450.
[6] Sojung Kim, Young-Jun Son, Ye Tian, Yi-Chang Chiu, C. Y. David
Yang, "Cognition-Based Hierarchical En Route Planning for MultiAgent Traffic Simulation", Expert Systems with Applications, pp., 2017, ISSN 09574174.
[7] Xiaoxu Zhang, Ying-Chang Liang, Jun Fang, 'Bayesian learning based multiuser detection for M2M communications with time-varying user activities", Communications (ICC) 2017 IEEE International Conference on, pp. 1-6,2017, ISSN 1938-1883
[8] Xenofon Fafoutis, Evgeny Tsimbalo, Robert Piechocki, 'Timing Channels in Bluetooth Low Energy", Communications Letters IEEE, vol. 20, pp. 1587-1590,2016, ISSN 1089-7798
[9] Paolo Attilio Pegoraro, Alessio Meloni, Luigi Atzori, Paolo Castello, Sara Sulis, "PMU-Based Distribution System State Estimation with Adaptive Accuracy Exploiting Local Decision Metrics and IoT Paradigm", Instrumentation and Measurement IEEE Transactions on, vol. 66, pp. 704-714,2017, ISSN 0018-9456

[10] Rosrio Valente, Chris De Ruijter, Daniel Vlasveld, Sybrand Van Der Zwaag, Pirn Groen, "Setup for EMI Shielding Effectiveness Tests of Electrically Conductive Polymer Composites at Frequencies up to 3.0 GHz", Access IEEE, vol. 5, pp. 16665-16675,2017, ISSN 2169-3536.
[11] Luca Mainetti, Vincenzo Mighali, Luigi Patrono, "A Software Architecture Enabling the Web of Things", Internet of Things Journal IEEE, vol. 2, pp. 445-454,2015, ISSN 2327-4662.
[12] Manuele Rusci, Davide Rossi, Michela Lecca, Massimo Gottardi,
Luca Benini, Elisabetta Farella, "Energy-efficient design of an alwayson smart visual trigger", Smart Cities Conference (ISC2) 2016 IEEE International, pp. 1-6,2016.
[13] Jol Toussaint, Nancy El Rachkidy, Alexandre Guitton, "Performance
analysis of the on-the-air activation in LoRaWAN", Information Technology Electronics and Mobile Communication Conference (IEMCON) 2016 IEEE 7th Annual, pp. 1-7, 2016.
[14] T. Gomes, F. Salgado, S. Pinto, J. Cabral, A. Tavares, 'Towards an FPGA-based network layer filter for the Internet of Things edge devices", Emerging Technologies and Factory Automation (ETFA) 2016 IEEE 21st International Conference on, pp. 1-4,2016.
[15] Mauro Biagi, Francesca Cuomo, Massimo Perri, Ahmad Irjoob, "A
Multi-Layer Parametric Approach to Maximize the Access Probability of Mobile Networks", Access IEEE, vol. 4, pp. 6692-6703,2016, ISSN 2169-3536.

4. SUMMARY OF THE INVENTION
In this work, the invention to make motorcycle riders aware of topographical deformities coming up ahead in the ride path is proposed. The proposed idea enables user to know about the presence of a pothole or a speed-breaker coming up ahead in his ride path so required speed reduction and manoeuvring of the motorcycle can be achieved safely. This invention is an attempt to make motorcycle riding safer. Knowledge of deformities beforehand resulting in proper speed reduction, accounts for lesser wear and tear of the motorcycle. Hence, the motorcycle requires lesser maintenance expenditure. The system logs exact GPS co-ordinates of deformities, which can be used by the concerned authorities to initiate repairs. An integrated speedometer in the android application allows for a unified system that can provide the real-time accurate speed fetched from the accelerometer of the smartphone.The main aim of this invention is to provide a safer riding experience to the motorcycle rider with ease of navigation. The proposed invention can not only be used by the motorcycle riders but can be used by motor vehicle users, public transport vehicles, private transportations like cabs etc.

CLAIMS:
We claim that:
(1) The invention is first of its kind setup to monitor, log and differentiate between deformities on roads.
(2) The proposed system is modular in design.
(3) The proposed system integrates its own android application that provides deformity updates.
(4) The proposed system can be deployed as kits in any existing or production ready motorcycle.
(5) The proposed system is capable of withstanding extreme environmental conditions.
(6) The proposed system is capable of detecting potholes with minimum width of 12cm.
(7) The proposed system is capable of detecting potholes with minimum depth of 6cm.
(8) The proposed system is light in weight and robust.
(9) The proposed system is capable of detecting speed-breakers with minimum height of 4cm.

(10) The prototype system provides precise speech and visual notification for deformities 50 to 100 metres before they are approached. The interval for notification depends on relative speed of the motorcycle.
(11) The system integrates 2 LEDS with blue and red colour for a flashing notification of speed-breaker and pothole respectively. This addition is for people who do not want to use Maps or don't want to mount their smartphone onto the motorcycle.