12th ITS European Congress, Strasbourg, France, 19-22 June 2017
Paper ID TP0927
Cooperative automation through the cloud: The CARMA project
Alan Stevens 1*, Mehrdad Dianati 2, Konstantinos Katsaros2, Chong Han2, Saber Fallah3,
Carsten Maple4, Frank McCullough5, Alexandros Mouzakitis5
1. TRL Ltd, United Kingdom, [email protected]
2. Institute for Communication Systems, University of Surrey, UK
3. Automotive Engineering Research Centre, University of Surrey, UK
4. Warwick Manufacturing Group, Warwick University, UK
5. Jaguar Land Rover, UK
Abstract
CARMA (Cloud-Assisted Real-time Methods for Autonomy) is a highly innovative and challenging
project which aims to develop and test cooperative automated driving technology, based on a
distributed control system. The approach is enabled by an ultra-low latency and highly reliable
cloud-based infrastructure accessed through 5G.
This paper describes the 3-tier distributed
computing architecture used in the project comprising the Vehicle, the Edge cloud and the Core.
It
describes a methodology to test a set of use cases representative of both urban and highway driving
and explores the key challenges in such an approach. The first technical challenge is the design of a
mobile edge cloud infrastructure that is able to support real-time and safety critical applications.
Another non-trivial problem is that of cyber-security for such a real-time cyber physical system.
Progress during the first year of this five year project is described.
Keywords:
C-ITS, 5G, Cloud Connectivity
Introduction
A combination of connected vehicles and automation is expected to significantly boost the
performance, safety, and reliability of next generation road vehicles. Co-operative automation is based
on vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) communications and the anticipated
availability and performance of 5th generation cellular networks (5G) is expected to provide exciting
new options beyond that provided by Dedicated Short-Range Communications based on the IEEE
802.11p protocol standard.
Cooperative automation through the cloud: The CARMA project
CARMA (Cloud-Assisted Real-time Methods for Autonomy) is a 5-year research project co-funded by
the UK’s Engineering and Physical Sciences Research Council (EPSRC) and Jaguar Land Rover under
a programme of five projects collectively called “Towards Autonomy - Smart and Connected Control”
where the automation focus is BASt Level 3+. The aim of CARMA is to develop and test cooperative
automated driving technology, based on a distributed control system, which is enabled by an ultra-low
latency and highly reliable cloud-based infrastructure accessed through 5G.
The remainder of this paper is structured as follows. The next section presents the 3-tier logical
architecture of CARMA and describes the key functionalities of the three main sub-systems. Then, the
future methodology is outlined. Next key use cases and applications are discussed followed by the
open issues, challenges and future work planned.
CARMA framework
The CARMA approach is that of a distributed control system, i.e., the executions of autonomous
functions are distributed between the on-board system and the cloud-based high performance shared
back-end system.
The system architecture is logically divided into a 3-tier distributed computing
system as illustrated in Figure 1, namely the CARMA Vehicle, the CARMA Edge and the CARMA
Core.
Figure 1 – CARMA 3-tier logical architecture
2
Cooperative automation through the cloud: The CARMA project
The on-board vehicle network connects various on-board sensors, infotainment equipment, on-board
embedded processors, HMI equipment, and actuators to apply control commands. The on-board
control components operate in cooperation with the edge controller assuring fault-tolerance of the
system in cases when the connection with the CARMA Edge is disrupted. Further, since vehicle safety
is of paramount importance, on-board controllers are also responsible to assess and potentially
override the remotely computed instructions (from cloud/edge) to ensure safety of the vehicle.
The CARMA Edge sub-system hosts off-board processes and information that require tight access
(low latency) with the vehicles. This will include information collected from around the vehicle and
processes that require cooperation with roadside equipment and other vehicles. This concept is
borrowed from ETSI Mobile Edge Computing (MEC) framework, where MEC servers can be installed
at Road Side Units or cell sites. In addition to Network Function Virtualisation (NFV) that will be
employed to support 5G communications, Connected and Autonomous Driving (CAD) specific
applications and functions will be off-loaded to assist with autonomous driving.
The cloud-based back-end system is based on a commercially available 3rd party server with a network
of high-performance shared processors. The back-end system also provides interfaces for a variety of
stakeholders that provide or use information from CARMA vehicles. This platform is expected to
enable highly autonomous (level 3) and fully autonomous (level 4) functions in a safe and cost
effective manner.
The key enabler of this system is a highly reliable and secure ultra-low-latency wireless
communication system among the vehicles, roadside equipment, and the cloud-based back-end. This is
envisaged to be provided through 5G communications and suitable security protocols.
Applications and use cases
CARMA aims to implement, assess and demonstrate a set of selected use cases representative of both
urban and highway driving.
3
Cooperative automation through the cloud: The CARMA project
Figure 2 – Example hierarchical task analysis of overtaking sub use case
To support the choice of specific use cases to exercise the CARMA architecture, a set of functions
(such as merging and leaving the highway, free driving, lane changing and overtaking, cooperative
driving/platooning, emergency braking, and obstacle avoidance) have been analysed using a
Hierarchical Task Analysis approach (see Figure 2).
Methodology for evaluation
The outline methodology is presented in Figure 2. The use cases will be analysed to understand the
requirements of the users (drivers) and the system requirements necessary to achieve these. These
use cases will then be mapped on to the CARMA logical architecture and implementation will lead to
a set technical requirements which are, in principle, testable.
4
Cooperative automation through the cloud: The CARMA project
Use Cases
User Requirements
System Requirements
Architecture 1
Architecture 2
Architecture N
Technical
Requirements
1
Technical
Requirements
2
Technical
Requirements
N
Research questions
Performance metrics
Experimental hypotheses
Experimental procedures (environments, design …)
Figure 3 – CARMA assessment methodology
CARMA will then develop a set of research questions and performance metrics and decide the
experimental domain in which these will be tested.
Potential experimental domains include,
modelling, bench testing, driving simulators, test tracks and (possibly) real road conditions. Each
sub-use case has a wide range of operation characteristics so a scenario approach will be used to
describe the conditions under which a use case may be executed. The conditions are described by a
series of Variables and the Variables are further specified by a series of Parameters. The Variables of
principal interest to CARMA may be grouped into those related to the following broad areas:
1. Subject vehicle
2. Driver of the subject vehicle
3. Road Infrastructure
4. Other road users
5. Communications capabilities
6. Environmental conditions
Key challenges and next steps
The CARMA project is highly innovative and challenging as designing a fault-tolerant distributed
cloud-based control system is non-trivial. The first technical challenge relates to the design of a mobile
edge cloud infrastructure that is able to support real-time and safety critical applications. This requires
an ultra-low-latency and highly reliable connectivity system both for V2V and V2I communications.
While LTE-V has been recently introduced for vehicular communications, with enhancements for V2V
communications through the side-link (device-to-device), Vehicle-to-Infrastructure is still based on
5
Cooperative automation through the cloud: The CARMA project
standard LTE/LTE-A system and there is still no native multicast/broadcast mechanism for message
dissemination, requiring the eMBMS Framework for distributing content of common interest to a
multitude of users in a spectrally efficient way. CARMA project will investigate advances in LTE-V
towards 5G communications with a focus on mobility management and other techniques to reduce
latency and increase reliability particularly on Vehicle-to-Cloud communications, with the edge being
the end-point most of the times. Further, due to the diversity of requirements from automotive
applications, network slicing may be sought to create end-to-end isolated networks with appropriate
configurations. Another non-trivial problem is that of security of real-time cyber physical systems. The
complexity and level of connectivity of the CARMA framework could potentially introduce a range of
risks and security threats, and whilst some analysis of security of wireless vehicle infrastructure exists,
much work is required to develop a rigorous threat-modelling framework that can be directly applied
to cloud-connected vehicles. The key research challenge will be the development and selection of
security techniques, protocols and standards that are compatible with the low-latency, distributed,
mobile, and heterogeneous nature of the system.
The CARMA project has made good progress in its first year of this five year project. The next step is
to carefully analyse the selected use cases and map them on to the logical architecture taking full
account of security issues.
Acknowledgements
This work was supported by Jaguar Land Rover and the UK-EPSRC grant EP/N01300X/1 as part of
the jointly funded Towards Autonomy: Smart and Connected Control (TASCC) Programme.
6