Report on Autonomous Cars

Introduction of Autonomous Cars

The automotive industry is advancing day by day. Introduction of new technologies has replaced the old fashion cars, as well as advanced features, has reduced the risk factor of on-road accidents. Automotive industries are specifically focused on the basic requirement of clients identified as "safety". For this purpose, companies had installed innovative features of detection and suggestions. Moderns cars are manufactured with the feature to detect objects and obstacles on the roadside during driving. Installed sensors and detection systems such as radar are programmed to process collected information and transmit it for the decision-making process. Such sensors equally facilitate vehicle drivers and driverless cars to increase safety and reduce accidental risk during the drive. The present work is aimed to study and discuss the use of electro-optic sensors and other multiple sensors in automobiles. In this work, secondary research data is collected from the research articles and other creditable sources to reach the conclusion. The research project is primarily focused on the multiple optical sensors data fusion for autonomous vehicles in our society.

Literature Review of Autonomous Cars

According to the research findings¸ electro-optic sensors are being in the use of automobiles to support them in handling uncertainties and traffic issues. Although, some sensors enable the drivers to have details about the multiple functions and alters about the vehicle. Information streams based on the multiple sensory systems of the vehicle provide the virtual environment and ensure real-time navigation systems in the highly intense traffic conditions. Following research findings, electro-optic sensors in automobiles are beneficial for real-time mapping. In the past few years, vehicles were not enough advanced to provide navigation systems to drivers. While now advanced cars and driverless cars are working on all functions in a better way with these sensor systems installed in various operations and systems of the vehicle. Cameras installed with these detection sensors have a feature to show the road maps and move cars at speed while detecting all obstacles and objects at roadsides. Installed electro-optic sensors in the vehicles work as edge detection, stereo image disparity, generation of electrical optical data, and blob detection. The fusion of this kind of collected information with the feature of GPS and AHRS in the vehicles reducing challenges for the automation and driverless car experiences (NEDOMA, et al., 2018).

 According to the research study, a parallel grayscale processing algorithm in the vehicles ensures the vision filtering process starting from image acquisition (using sensors and cameras) and ends at the stage of data extraction and object identification. In general, researchers claim that data obtained from the sensors and cameras are processed in such a way that provides a clear picture of available challenge or problem. While other sensors installed with the specific programs respond automatically towards these identified problems by using already provided or added applicable suggestions in the vehicle. For instance, while moving on a road a vehicle acquires images that are encoded to the system for the vision filtering process. Thus, at the end of this process, the vehicle identifies the object and move aside from that object to avoid an accident. However, the whole process just takes very minimum time duration because of the efficiency of sensors and systems installed in the automobile.

In accordance with the research findings, sensors also support inter-vehicle communication systems. A complex driving environment generates challenging situations for the driver. To meet these complex and unique challenges in an appropriate manner, drivers are required to interact with the sensors in the vehicles. Sensors collect data and push it to the driver to make their driving safer and more reliable. However, in the main process vehicles require internal vehicle communication systems that work with the onboard sensors such as electro-optic sensors, lidar, cameras, and radars. Following the information presented in the research article, vehicles with these advanced sensor systems should not over-rely on vehicle environment perception systems and local data provided by the vehicle sensors. Instead, vehicles should also collect data from the network information from remote vehicles to reduce accident risk and other uncertain hazards. A perfect combination of local environmental data and vehicle sensor systems can reduce such risk and make a drive of driverless cars and other cars more reliable. Researchers identified that CAN, DSRC, GPS, Radar, and other electro-optic sensors are reliable technologies to be used in automobiles.

Summarizing the research findings of Haroglu, Powell, and Seyam (2017), the automotive industry is working on the continuous improvement concept to introduce innovative solutions to its targeted customers with a higher focus on competitive advantage using technologies. A new advancement in this industry is polymer optical fiber (POF) sensors. These sensors have a feature to collect data and transmit them to the relevant attached devices or machines. In-vehicle, such sensors are in use with high bandwidth and immune to electromagnetic interference (Haroglu, Powell, & Seyam, 2017).

The researchers Arjun Balasuriya, Zhen Jia and Subhash Challa have provided details regarding the vision-based data fusion for autonomous vehicles. In this research study, the algorithm is proposed through which autonomous vehicles can perform object tracking. For evaluating the speed of the object the algorithm utilizes the information which the autonomous vehicle’s sensors capture. The algorithm also utilizes the information captured by the inertial motion sensors and the cameras which are placed in the autonomous vehicle. The findings of the studies have provided information regarding the IMM tracking algorithm and how much effective this algorithm is in tracking the object (Jia, Balasuriya, & Challa, Sensor fusion-based visual target tracking for autonomous vehicles with the out-of-sequence measurements solution, 2008)

The researchers Zhen Jia, Arjun Balasuriya, and Subhash Challa have provided details regarding the sensor fusion based target tracking for autonomous vehicles. In this research study, the algorithm is proposed through which autonomous vehicles can perform target tracking. For evaluating the velocity of the object the algorithm utilizes the information which the autonomous vehicle’s sensors capture. The algorithm also utilizes the information captured by the inertial motion sensors and the cameras which are placed in the autonomous vehicle. The findings of the studies have provided information regarding the visual target tracking algorithm and how effective this algorithm is in target tracking (Jia, Balasuriya, & Challa, Vision based data fusion for autonomous vehicles target tracking using interacting multiple dynamic models, 2008).

The research study conducted by Hitesh Laware, S. Reza Zekavat, David Castanon and Sourav Chakraborty has provided deep insights about wireless technologies and data fusion for localization for autonomous vehicles. According to the researchers, the navigation systems in the vehicles are utilized usually for 3 purposes which include navigation, positing and routing. Researchers have stated that GPS has its limitations and due to its limitations, it cannot perform all the tasks. In this study, a new system is proposed which has utilized different wireless technologies for reliable navigation. The proposed system aims to enhance the security and safety of the vehicles (Chakraborty, Laware, Castanon, & Zekavat, 2016).

The study conducted by C.J. Harris and R.J. Walker has discussed the importance of a multi-sensor fusion system in autonomous vehicles. The technology is becoming advanced day by day and the vehicles today are utilizing the latest ways to enhance security and safety. The traditional methods are not as robust as the latest technological advancements. In this study, MSDF has shown that such approaches enhance the system robustness and reduces the chances of errors and failures. The organizations have to utilize the latest technologies to enhance system robustness (R.J.Walker & C.J.Harris, 1993).

The study conducted by Mohammad K. Al-Sharman, Bara J.Emran, Mohammad A.Jaradat, and Homayoun Najjaran has discussed the importance of multi-sensor fusion system. According to the researchers, the traditional methods are not as robust as the latest technological advancements. In this research study, the algorithm is proposed through which autonomous vehicles landing performance is measured. The findings of the studies have provided information regarding the proposed algorithm and how effective this algorithm is (K.Al-Sharman, J.Emran, A.Jaradat, & Najjaran, 2018).

 As described by Raol & Girija (2002), the increasing importance is being increased by multi-sensor information fusion into modern technologies. Furthermore, the information fusion technique’s merit is basically the fusion on the level of measurement which is quite straightforward. For the possible application to the data fusion sensor process, the square root information filter is studied. To deal with the problem, the decentralized square root information filter was proposed by the researcher of this study. To attain the accuracy numerically, the stability, as well as the reliability of the fusion algorithm, the implementation of the square root filter, would be very useful. It provides the meaningful results of the validation of a decentralized square root information filter by using the simulated data (Raol & Girija, 2002).

            As stated by Zanchin, Adamshuk, Santos, & Collazos (2017), the carmakers are toward the car's new generation which tends to become more autonomous as possible as without having or being the driving intervention such as the intelligent transportation systems part. The autonomous vehicles have the enhancing ability to drive the vehicle in an automated way, high accuracy to reduce the possibilities of accidents as well as the congestion of the traffic considering the performance, safety as well as improvement in the comfort of driving. The authors have presented the information on autonomous cars in the form of discussion in this literature which is on the classification as well as instrumentation of the autonomous cars for the community that is interested to understand the autonomous vehicle field deeply.

            Furthermore, the autonomous cars and vehicle definitions are also introduced by Zanchin, Adamshuk, Santos, & Collazos (2017) which is related to the society of automotive engineering (SAE international) as well as NHTSA (United states department of the national highway of traffic safety administration of transportation). Moreover, the autonomous cars definition order as well as categorize the autonomy of autonomous cars levels. Although, it also symbolizes through very complicated capabilities as well as functions for every described level by SAE as well as the higher vehicle autonomy level and the less dependence on the intervention of humans. It was employed by the sensors used in the autonomous cars and vehicles from or since the sensors unidimensional such as scanner, radar up to the camera as well as LIDAR. Furthermore, it was also discussed in the journal by the researchers that the fusion sensors are very essential as well as significant elements of the autonomous cars which gives the environmental wide vision and clear insight in which it inserts the cars to make effective decisions to accept the conditions of driveability accordingly because it also generates the feeling and increases the desire to reach at the desired places immediately (Zanchin, Adamshuk, Santos, & Collazos, 2017).

Results, Project objectives, the performance of technology

            The unique challenges are posed by the autonomous driving to the perception of the environment of a vehicle because of the complicated environment where the autonomous vehicles should be identified as well as they interact or connect with the surrounding vehicles simultaneously. As this section of the research study is describing the project objectives as well as the results of the performance of the technology, the researchers have discussed their proposed approach also. The researchers of this study have developed an inter-vehicle communication system for getting high technology performance of the technology practically along with the reliability as well as the simplicity of the practicality of the technology. The important data synchronization issue is dealt with based on a dead reckoning strategy in the presence of miss detections. For the object matching in real traffic on the road, an IMM-based track association is creatively presented by the use of SMHT (IMM-SMHT) as well as achieving the exceptional matching of results of the important practical implications. Furthermore, the data properties are also carefully considered by the researchers in the difficult statistical sense through making the scalability possible for the application of the diverse statistical techniques directly. So, the important contributions of the autonomous cars study consist of such aspects which are given below.

·         The results of this study are showing that the researchers designed the object matching algorithm to provide the scalable as well as reliable statistical results. As the ubiquitous solution, the IMM-SMHT is presented for several kinds of vehicle driving scenarios. Furthermore, it can also apply several advanced sensor fusion approaches to improve the performance of the system.

·         The system of the intervehicle communication system is developed on the modules of the current hardware of high practicability such as DSRC transceiver, the automotive radar as well as low-cost GPS. It has also earned a high experience in every part of the hardware module. So, the reliability and simplicity are guaranteed as well as it can also directly extend the system for more configurations generally.

·         The problems or challenges such as data synchronization and the data drops are managed reasonably in the design of the algorithm. While the properties of the collected data statistically are better evaluated. Furthermore, it also employs the logic layer for the countermeasure of some known artificially introduced physics.

Challenges or Limitations of Autonomous Cars

            The researcher at the time of conducting the research maintains the balance among the implementation as well as performance complexity in the schedule. It handles the issues in the synchronization, delay of the data as well as dropout of data. Furthermore, the rigorous studies based on advanced theory can also be helpful on some of the auto-correlation challenges, as well as the drifting issue of the extended object is carried out independently. The problem in the object matching technique is very challenging as well as it also requires improved statistical processing by comparing several systems based on the camera.

            The challenging part of the object matching lies apparently in the reliable and appropriate association of the two data sets that DSRC data from the remote vehicles which were observed or identified by radar. It shall also pick the particular common information to serve as the cues of the association of data such as the kinematic information of velocity, acceleration or the position of the autonomous car as well as the information of categories such as feature, speed profile, and size (Yuan, et al., 2017). Thus, the researchers of this study have only used the information of the position for the purpose of object matching to fully exploit the proposed system potentially. Furthermore, the researchers of this study tried to attain the goal by using recent information.

            As the list of detect objects by object matching algorithm, the candidates of radar are output with the confidence. An internal object ID, the radial velocity and angle or depth measurement or depth are contained by every object usually based on the doppler effect. Furthermore, the rough dimension estimations based on L-shape reflecting points can be given by improved radar systems. Therefore, dimension estimation is not completely consistent because of the corner of a vehicle or the radar. The extended object tracking approach is also anticipated in this matter.

            Furthermore, the technology is more sensitive to several problems such as communication bandwidth and data dropouts or the miss detections as well as time synchronization.

Conclusion of Autonomous Cars

It is concluded that the electro-optic sensors are being in the use of automobiles to support them in handling uncertainties and traffic issues. Cameras installed with these detection sensors have a feature to show the road maps and move cars at speed while detecting all obstacles and objects at roadsides. A complex driving environment generates challenging situations for the driver. To meet these complex and unique challenges in an appropriate manner, drivers are required to interact with the sensors in the vehicles. The algorithm is proposed through which autonomous vehicles can perform object tracking. For evaluating the speed of the object, the algorithm utilizes the information which the autonomous vehicle’s sensors capture. The navigation systems in the vehicles are utilized usually for 3 purposes which include navigation, positing and routing. The traditional methods are not as robust as the latest technological advancements. For the possible application to the data fusion sensor process, the square root information filter is studied. The unique challenges are posed by the autonomous driving to the perception of the environment of a vehicle because of the complicated environment where the autonomous vehicles should be identified as well as they interact or connect with the surrounding vehicles As the ubiquitous solution, the IMM-SMHT is presented for several kinds of vehicle driving scenarios. It has also earned a high experience in every part of the hardware module. Furthermore, the rigorous studies based on advanced theory can also be helpful on some of the auto-correlation challenges, as well as the drifting issue of the extended object is carried out independently.

 References of Autonomous Cars

Chakraborty, S., Laware, H., Castanon, D., & Zekavat, S. R. (2016). High precision localization for autonomous vehicles via multiple sensors, data fusion and novel wireless technologies. 2016 IEEE 7th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON), 1-9.

Chakraborty, S., Laware, H., Castanon, D., & Zekavat, S. R. (2016). High precision localization for autonomous vehicles via multiple sensors, data fusion and novel wireless technologies. 2016 IEEE 7th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON). doi:10.1109/UEMCON.2016.7777799

Haroglu, D., Powell, N., & Seyam, A.-F. M. (2017). The response of polymer optical fiber (POF) to bending and axial tension for the application of a POF sensor for automotive seat occupancy sensing. The Journal of The Textile Institute, 108(1), 132-139.

Jia, Z., Balasuriya, A., & Challa, S. (2008). Sensor fusion-based visual target tracking for autonomous vehicles with the out-of-sequence measurements solution. Robotics and Autonomous Systems, 56(2), 157-176.

Jia, Z., Balasuriya, A., & Challa, S. (2008). Vision based data fusion for autonomous vehicles target tracking using interacting multiple dynamic models. Computer Vision and Image Understanding, 109(1), 1-21.

K.Al-Sharman, M., J.Emran, B., A.Jaradat, M., & Najjaran, H. (2018). Precision landing using an adaptive fuzzy multi-sensor data fusion architecture. Applied Soft Computing, 69, 149-164.

NEDOMA, J., FAJKUS, M., KAHANKOVA, R., MARTINEK, R., DVORSKY, M., VANUS, J., . . . CVEJN, D. (2018). Fiber-optic interferometric sensor for monitoring automobile and rail traffic. Turkish Journal of Electrical Engineering & Computer Sciences, 26, 2986-2995.

R.J.Walker, & C.J.Harris. (1993). A Multi-Sensor Fusion System for a Laboratory Based Autonomous Vehicle. IFAC Proceedings Volumes, 26(1), 107-112.

Raol, J., & Girija, G. (2002). Sensor data fusion algorithms using square-root information filtering. IEE Proceedings - Radar, Sonar and Navigation, 89 – 96. doi:10.1049/ip-rsn:20020128

Yuan, T., Krishnan, K., Chen, Q., Breu, J., Roth, T. B., Duraisamy, B., . . . Gern, A. (2017). Object Matching for Inter-Vehicle Communication Systems—An IMM-Based Track Association Approach With Sequential Multiple Hypothesis Test . IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 3501-3512.

Zanchin, B. C., Adamshuk, R., Santos, M. M., & Collazos, K. S. (2017). On the instrumentation and classification of autonomous cars. 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC). doi:10.1109/SMC.2017.8123022