Automotive Communication and Control Trends

Control Requirements
Control system by itself implies several data streams with hard synchronization in time. All components of the communication system must support this mode and flexibly connect control system sensors, actuators, and control units.

Modern trends in the industry show that the number of local control systems constantly grows. For instance, in the latest models BMW offers nine main electronic control systems. For implementation of more complex options, closed control loops connect with each other by transmitters or communication system bridges.

Such communication is determined not just by interconnection of a multitude of sensors and actuators in different control systems. Architectural communication is also defined by outputs of one system on parameters of other control systems. These factors impose stringent requirements on communication channel data rates.

As estimated by the FlexRay consortium, rates must be up to 10 Mbit/sec. At the same time new systems must support time-determined control. They must provide the possibility of synchronizing both parameters of controlled objects (units) and parameters of vehicle control systems. This process is carried out by the principle of a "digital snapshot" of a controlled object.

Special attention in the list of requirements is given to reconfigurability and fault-tolerance of the systems. Focus is placed at composability and unification of bus systems for software reuse and decrease of costs. This is a global trend: In 2004 in Japan a JASPAR (Japan Automotive Software Platform and Architecture) alliance of major automotive manufacturers was formed for advancing unified solutions in vehicle software and electronics.

Data Architecture Influences on an Automobile
Data stream architecture increasingly influences the whole automobile, its characteristics, and options. The architecture also determines the topology of the power supply network, placement and characteristics of mechatronic units, etc. It also determines functionality and requirements of vehicle systems.

Traditional engineering approach implies that local control systems are connected only by a communication channel and its components. At the same time, every new electronic control system is added into an already formed architecture, which leads to the problem of the new system's proper functioning in an automobile.

From the control point of view, the main technical difficulty is the local system is affected by all other control systems of an automobile dynamically in time. If this influence is not taken into consideration, then accuracy and efficiency of the system decrease. As a result, significant technical resources, of both channels and computing, are spent on conflict resolution, and splicing and connecting systems with each other.

Accurate representation of all processes is changed by selection of the appropriate algorithm for control system and its criteria. Thus, vehicle control "by systems" complicates its architecture and is not flexible. From the system development point of view it is very difficult to integrate new systems into the automobile and achieve stable control in them except for objective technical limitations.

Existing Architectures and Communication Standards
Existing architectures implement multiplexed communication systems. The widely used CAN standard utilizes, for instance, communication protocol by tasks. This technology partially solves the problem of synchronization of system functioning in time and by parameters. Simultaneously channel loading in CAN greatly increases with channel-related tasks that lead to sharp increase of frequencies and transmission rates.

The Society of Automotive Engineers (SAE) generalized and formed technical requirements of automotive engineers for communication systems. Basic parameters of communication systems and SAE requirements are shown in below.

View Full Size Table

Problems and Trends
Prospective requirements for communication systems developed by major alliances of manufacturers are in many respects defined by practical experience of vehicle development. But at the same time, they emphasize the need for more effective technical solutions.

For example, technical analysis of the CAN protocol information density shows 28% of service information takes up to approximately 72% of a message (figure below). This means that at maximum speed of 2 Mbit/sec without hindrances, a message transmits up to 560 kbits of control information. Further increase of speed leads to proportional slight increase of "useful information" and is limited to physical ability of transmission.

View Full Size Image

The same tendency of searching for effective solutions is found when trying to replace wire systems with optical and wireless communication systems. The main problem of the first solution is its durability and cost, the second solution requires increasing of parameters and dependable protection from hindrances (interference). Solutions that utilize time-triggered protocols (TTP) are the most suited for control principles. An objective analysis shows TTP-based systems' high (up to 90 %) information density and dependability, but shortcomings include development complexity and lack of flexibility for automotive applications.