Self-Driving Cars and Human Factors - Research Paper

Introduction

Autonomous vehicles are not only experimental but are here to stay. The fact that the vehicles are self-driving still requires care and caution from even the most reckless driver. Autonomous vehicle has been certified as safe but the safety of the autonomous vehicle is determined by the accuracy of the algorithms designed for path planning and decision making. In the urban areas, path planning is easy because most of the building and structures are permanent and so the autonomous vehicle can map their positions and store the data in its memory for future referrals. The path planning is also a strategy used by the self-driving cars to help them find a safe and convenient route by calculating the number of vehicles, the traffics and the probability to delays being in a route (Chen et al. 2019, p. 33). Most self-ruling cars are fitted with the other algorithms to help them find the most economical route among many tours and would combine safety, convenience, and costs in its route determination equation.

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Save 25%

The safety of the self-driving cars is therefore determined by the costs, convenience, and safety. Finding a geometrics path from one point to another involves determining the path's feasible. The car must maneuver its routes, based on the position of the vehicle and its speed on the road. Whenever the vehicles are planning its maneuver, it is important to note that 94% of the accidents are caused by the drivers while only 2% of the accidents are caused by the vehicles and 2% caused by the unknown critical reasons. Accidents may also be due to recognition error, performance error and or decision error. Never the less, tires related accidents, brakes related accident, steering or suspension related accidents or transmission related accidents are also common in varying frequencies (NetSA, 2015, p. 3).

The trajectory of the vehicles of the sequences of the states that the vehicles visits can be parameterized by the time and velocity. The trajectory can cause an accident if path trajectory planning is poorly calculated which attributes the accidents to the vehicle's software. Self-driving cars should generate trajectory in real time that is why real-time planning of the movement of the vehicles from one point to another is important. Any error is trajectory generation could lead to accidents as the delays in trajectory generation might cause the vehicles to miss is kinematic limits as determined by the dynamics. the navigation mode may constraint the movement of the vehicles if the software is faulty.

The vehicle's safety is driven by how fast the vehicles can consciously search four space and deviating path. Path planning involves transforming real-life physical environments into a series of digital configuration for planning. The vehicles identify obstacles, edges, closets points, and paths and map their sizes and potential impacts which are communicated to the vehicle's navigation systems. The algorithm can generate paths and maximize the distances between the vehicle's other obstacles in the immediate environment which might include vehicles barriers pedestrians. The risk and feasibility of the paths are calculated by taking into consideration the presence of other obstacles in the road and other road boundaries. According to Standard (2017, p. 3), digital communication to vehicles at an intersection will require a significant change of the ingratitude as vehicles have to decide between the static signs or variable traffic light. The cost maps algorithms, slate lattices algorithms and driving corridors algorithms are all used in reducing the possibility of risk causing factors in the road. The most recent algorithms are the Simultaneous Location and Mapping (SLAM) models which is effective in detecting objects in the paths and near the self-driving cars both mobile object and stationary objects.

Vehicle Index Vehicle location from the start of the road (in meters) Calculated distance from the emergency initiator Vehicle A (in meters)

Figure SEQ Figure \* ARABIC 1: Neighbor Location Table. Adapted from Kaur and Kang (2016)

Self-Driving cars can determine the distances without accidents or safe distances. For example, the best distances without accidents may be 0.38 miles, the current distance may be 0.42 while the time between the car and accident causing objects may be 32 minutes, and 31 seconds which will be fed back into the car may halt based on the feedback.

According to Chen (2016), the cognitive model can be applied in modeling the predictive be autonomous vehicles use of maneuverable objects. For example, the cognitive model with attention (CAM) is used to process the road images using multiple convolutional neural networks to help in simulating the function of the human visual cortex. The cognitive map containing the vehicles' status, the autonomous vehicle navigational input as well as the perception results is built and the resulting information is stored in the cognitive map and used to make driving decisions.

Human factors in autonomous vehicles only come into play at the vehicle programming stage. a completely autonomous vehicle uses a wide range of algorisms which when combined with the vehicular ad hoc network (VANET) can significantly reduce autonomous vehicles accidents. The VANET allows autonomous vehicles to communicate autonomously. The closets attempt to enable autonomous communication by vehicles was through the Controller Area Network bus (CAN BUS) proposed by Currie (2019). The data shared include data on the traffic conditions, the situation, and condition of the vehicle and other internal data that can be used by the vehicles to avoid traffic jams or prevent them from causing accidents. If the vehicles can communicate autonomously in real times, the other vehicles can easily make driving decisions on factors such as speed and direction. Never the less, the fact that the vehicles are anonymous can easily lead to an accident which is why programmers must contend with the balance between anonymity and privacy of the vehicles. To improve safety, the vehicle data should be publicly available but this will mean a breach of privacy. However, the main limitation of the VANET is that it is quite challenging to achieve successful multi-hop data delivery especially when frequents disconnection is required in the high mobility.

Methodology and Plan

The iterative software development life cycles will be adopted. under the iterative mode, the researcher will first raft the requirements to the functional parts with the goals of expanding then second later. The processes may be repetitive as the researcher will seek to refine each part to develop a more refined version. Every iteration will last only one week. The researcher will thereof stay with analyzing the systems requirements, then design the VANET systems, code the systems, test the systems and later implement. In case the software will not pass the alpha and beta testing stage, the whole iterative will be started again. Below is a framework of the

Figure SEQ Figure \* ARABIC 2: SDLC Iterative Model. Own impression

Iterative model is chosen because it allows for rapid development of the function at the beginning of the project. It also allows for the parallel development of parts. The fact that the iterative model is progressive and measurable makes it cost effective. The development can also manage the risk that would lead to cost overruns or delays. For example, the shooter the iteration the easier it becomes to debugging (Liu, and Yan). The developer can define and solve risk or problems in the early phases to prevent the problem from occurring in the latest phases of the projects. Finally, the model makes software development easy because the approach is flexible and allows the developer to make changes are at the stage of the software development lifecycle

Task Plan

Use A systematic review protocol PRISMA to search for peer-reviewed journals that meet the criteria

Conducting a literature review. Systematic analysis and meta-analysis will be usedDevelop a VANET prototype for inter-vehicle communication. Carry out systems requirement analysis, followed by system design, system development and, coding and testing before deployment.

Write a discussion of the research findings and prototype presentation

Provide recommendations for future research.

References

Currie, R. (2019). Developments in Car Hacking GIAC. (GSEC) Gold Certification. SANS Institute InfoSec Reading Room.

Kaur, G. and Kang, S. (2016). Technique to control Data Dissemination and to support data accessibility in Meagerly Connected Vehicles in Vehicular Ad-Hoc Networks (VANETS). International Journal of Advanced Research in Computer Science, 7(6).

Liu, X. and Yan, G. (2016). Analytically modeling data dissemination in vehicular ad hoc networks. Ad Hoc Networks, 52, pp.17-27.

Maynard, T., Beecroft, N. and Gonzalez, S. (2014). Autonomous vehicles handing over control: opportunities and risks for insurance. Lloyd's.

Blanco, M., Atwood, J., Vasquez, H. M., Trimble, T. E., Fitchett, V. L., Radlbeck, J., ... & Morgan, J. F. (2015, August). Human factors evaluation of level 2 and level 3 automated driving concepts. (Report No. DOT HS 812 182). Washington, DC: National Highway Traffic Safety Administration.