Development Of Self-driving Vehicles Research Articles

Decentralised cooperative cruising of autonomous ride-sourcing fleets

As transportation network companies and automobile manufacturers continue to invest in the development of self-driving vehicles, it can be expected that autonomous taxi (a-taxi) fleets will become a major component of on-demand transport services in the foreseeable future. The majority of existing automated fleet management systems focus on central dispatch strategies that rely on real-time information and communication. This paper proposes a novel decentralised cooperative cruising method for offline operation of a-taxi fleets, which serves as a contingency plan during a full communication shutdown. The proposed method acts as an emergency plan for the system to continue serving passengers with the objective of maximising the total number of served passengers by the fleet. The method uses historical trip data to estimate PageRank centralities of roads as a proxy of long-term likelihood of finding waiting passengers over a series of trips. The proposed method uses this metric to (i) compute weighted shortest paths for vacant a-taxi cruising route planning, and (ii) partition the network into homogeneous regions for effective cruising destination choice (mission planning). The movements of vacant a-taxis between regions are modelled as a Markov chain such that a transition probability matrix is computed to achieve the optimal spatial distribution of vacant a-taxis to maximise the total expected pick-ups by the fleet, estimated based on bilateral meeting functions. Compared to benchmark strategies which select destinations randomly and cruise along the shortest travel time path, the proposed method shows significant improvements in service performances for different fleet sizes.

BUILDING OPENDRIVE MODEL FROM MOBILE MAPPING DATA

Abstract. The development of self-driving vehicles has been efficiently accelerated by applying computer simulations. These environments can achieve the best result only if reality has been digitized: highly accurate and detailed maps are required. The high-definition (HD) maps aim to fulfill these requests. Companies and academia are seeking adequate technology, where OpenDRIVE standard seems to win this race.A software was developed to create high-definition road models for self-driving car testing purposes, which can also be used in driving simulators, for example, for testing vehicle prototypes.The data of a sample area was received located in northern Budapest, which included inconsistent information about the lanes of the road network. The top-priority task was to build a highly structured, consistent model satisfying this open standard, an increasingly popular modeling format to describe the environment for various simulating software. In order to do that, the relevant information had to be extracted from an AutoCAD file. The result was a collection of points that was added to a software module written in C++. The code has built a structure from those points, including determining connections between lanes, building roads from the lanes, determining connections between roads, adding road marks, and building the final OpenDRIVE model.Two different sets of data have been used thus far in the software: a circa 8 km long road section and a separate junction – meaning about 20,000 points to process. Two independent models were created; both were tested in various automotive simulators – including IPG CarMaker, probably the most well-known simulating tool used in car manufacturing – which showed that both are indeed correct. The overall time required to build those models from simple text files to well-structured OpenDRIVE models is no more than two minutes, while before the software development – when all of those aforementioned processes were done by humans – it required several weeks, if not months.The software is already capable of building basic road models but is nowhere near finished. Numerous other features can be added to it as well, for instance, traffic signs and signals and environmental elements such as roadside objects, trees, and buildings.

Are Self-Driving Vehicles Ready to Launch? An Insight into Steering Control in Autonomous Self-Driving Vehicles

In the last couple of years, academia-industry collaboration has demonstrated rapid advancements in the development of self-driving vehicles. Since it is evident that self-driving vehicles are going to reshape the traditional transportation systems in near future through enhancement in safe and smart mobility, motion control in self-driving vehicles while performing driving tasks in a dynamic road environment is still a challenging task. In this regard, we present a comprehensive study considering the evolution of steering control methods for self-driving vehicles. Initially, we discussed an insight into the traditional steering systems of the vehicles. To the best of our knowledge, currently, there is no taxonomy available, which elaborates steering control methods for self-driving vehicles. In this paper, we present a novel taxonomy including different steering control methods which are categorized into deterministic and heuristic steering control methods. Concurrently, the abovementioned techniques are critically reviewed elaborating their strengths and limitations. Based on the analysis, key challenges/research gaps in existing steering control methods along with the possible solutions have been briefly discussed to improve the effectiveness of the steering system of self-driving vehicles.

A Standard Driven Software Architecture for Fully Autonomous Vehicles

The development of self-driving vehicles is often regarded as adding a layer of intelligence on top of classic vehicle platforms. However, the amount of software needed to reach autonomy will exceed the software deployed for operation of normal vehicles. As complexity increases, the demand for proper structure also increases. Moreover, the shift from open, deterministic components to more opaque, probabilistic software components raises new challenges for system designers. In this paper we introduce a functional software architecture for fully autonomous vehicles aimed to standardise and ease the development process. Existing literature presents past experiments with autonomous driving or implementations specific to limited domains (e.g. winning a competition). The architectural solutions are often an after-math of building or evolving an autonomous vehicle and not the result of a clear software development life-cycle. A major issue of this approach is that requirements cannot be traced with respect to functional components and several components group most functionality. Therefore, it is often difficult to adopt the proposals. In this paper we take a prescriptive approach starting with requirements from a widely adopted automotive standard. We follow a clear software engineering process, specific to the automotive industry. During the design process, we make extensive use of robotic architectures – which seem to be often ignored by automotive software engineers – to support standard driven requirements.

Harnessing Uncertainty in Autonomous Vehicle Safety

Safely developing self-driving vehicles presents technical challenges. Among the key technical challenges are how to confidently demonstrate the safety of a self-driving vehicle when the number of permutations of operating conditions, scenarios, system inputs, etc. are complex, uncertain and potentially limitless. This paper provides a broad survey of the various types of uncertainty in the development of self-driving vehicles and outlines several possible strategies for handling uncertainty. Advantages and challenges of different approaches, including qualitative and quantitative methods, are also discussed.

Customer Preferences and Implicit Tradeoffs in Accident Scenarios for Self-Driving Vehicle Algorithms

The development of self-driving vehicles is proceeding rapidly and with significant investment of resources. However, a full-scale deployment is not imminent. Among the challenges self-driving vehicles are facing, they will have to navigate complex ethical challenges. The algorithms governing their behavior will have to decide how to steer them in situations where accidents cannot be avoided. In some of these situations they will have to decide which of several potential parties to injure in the process. We investigate the preferences of Swiss customers for this decision by forcing a selection between simplified scenarios where a given number of car passengers or a given number of pedestrians will be killed in the accident. Both passengers and pedestrians can be adults or children. The passengers are explicitly identified as the respondent themselves and their family. While children are implicitly valued higher than adults, Swiss customers value passengers and pedestrians implicitly roughly equally, and assign increasingly higher marginal values to additional people, both passengers and pedestrians. These results seem to partially contradict similar studies conducted in other countries and recent statements by automotive companies, potentially indicating the need to adapt both corporate communications and steering algorithms in different geographies.