Tezislər / Theses

THE 3
rd
 INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS 
dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
198
As navigation sensors may create real-time data streams at high data 
rates (typically in the gigabits per second), cars can't send all of their sensor 
data within the available bandwidth. One major finding is that different types 
of data collected by sensors are of varying relevance. Carcel takes 
advantage of this by instructing the cloud to seek information at a higher 
resolution from only the most critical parts of the environment. These 
requests are sent to the car via the cloud's request module. It analyzes 
sensor data in aggregate, represented using the Octree data structure, and 
creates requests for different locations at specific resolutions depending on 
sensor data lacunae. The receiver module feeds sensor data to the planner 
module, which collects data from different autonomous vehicle and static 
road-side infrastructure sensors. The receiver module on each autonomous 
vehicle registers requests and advises the transmitter to transmit a higher 
proportion of sensor data from the requested locations at the appropriate 
resolution. 
[2]
In overall, first we create a real-time connection pool between each 
endpoint of the agents which generates multitude of data points to gather 
information. We have applied multi-robot control systems over the 
autonomous cars for simulating more practically and give solid output from 
our hypothesis. That is why, our implementation on autonomous vehicles 
consists of hardware sensor logging, data transmission and data receiving 
from cloud platform. Additionally, the methodology contains model training 
under the support of cloud infrastructure with GPU acceleration. 
All things considered; the research paper suggests that cloud computing 
is increasingly in high priority to rocket the efficiency of multi-robotics 
systems and autonomous vehicles. In this paper, we have provided different 
types of approaches and solutions with different methodologies by other 
researchers as well as depicted some algorithmic scheme and diagrams to 
illustrate more visually the architecture which is intended. Utilizing cloud 
resources in equivalent load balancing techniques we can conclude with 
better design perspective which is claimed mainly in our paper. Thus, many 
agents that followed by cloud platform and were connected to the remote 
server can transfer data and then benefit from the common data storage and 
predictions in the cloud system.
References 
[1] S. Liu, J. Tang, C. Wang, Q. Wang, and Jean-Luc Gaudiot, (2017) “Implementing a Cloud 
Platform for Autonomous Driving”, (pp 1-8) Fellow, IEEE 
[2] J. Levinson, J. Askeland, J. Becker, J. Dolson, D. Held, S. Kammel, J. Kolter, D. Langer, 
O. Pink, V. Pratt, M. Sokolsky, G. Stanek, D. Stavens, A. Teichman, M. Werling, and S. 
Thrun. (2011) “Towards fully autonomous driving: Systems and algorithms”. (pp 163-168) 

---