Image Processing Reference
In-Depth Information
The error recovery time, defined as the time from detecting an error until the possible start of a new
frame, is - bit times. CAN possesses some fault-confinement mechanisms aimed at identifying
permanent failures due to hardware dysfunctioning at the level of the microcontroller, communi-
cation controller, or physical layer. The scheme is based on error counters that are increased and
decreased according to particular events (e.g., successful reception of a frame, reception of a cor-
rupted frame, etc.). The relevance of the algorithms involved is questionable (see Ref. []) but the
main drawback is that a node has to diagnose itself, which can lead to the nondetection of some crit-
ical errors. For instance, a faulty oscillator can cause a node to transmit continuously a dominant bit,
which is one manifestation of the “babbling idiot” fault, see Ref. []. Furthermore, other faults such
as the partitioning of the network into several subnetworks may prevent all nodes from communicat-
ing due to bad signal reflection at the extremities. Without additional fault-tolerance facilities, CAN
is not suited for safety-critical applications such as some future X-by-Wire systems. For instance, a
single node can perturb the functioning of the whole network by sending messages outside their
specification (i.e., length and period of the frames). Many mechanisms were proposed for increasing
the dependability of CAN-based networks (see Ref. [] for an excellent survey), if each proposal
solves a particular problem, they have not necessarily been conceived to be combined. Furthermore,
the fault-hypotheses used in the design of theses mechanisms are not necessarily the same and the
interactions between them remain to be studied in a formal way.
The CAN standard only defines the physical layer and Data Link layer (DLL). Several higher level
protocols have been proposed, for instance, for standardizing startup procedures, implementing data
segmentation or sending periodic messages (see OSEK/VDX and AUTOSAR in Section .). Other
higher-level protocols standardize the content of messages to ease the interoperability between ECUs.
hisisthecaseforJwhichisused,forinstance,inScania'strucksandbuses[].
13.2.1.2 Vehicle Area Network
Vehicle area network (VAN) (see Ref. []) is very similar to CAN (e.g., frame format and data rate)
but possesses some additional or different features that are advantageous from a technical point of
view (e.g., no need for bit-stuffing, in-frame response: a node being asked for data answers in the same
frame that contained the request). VAN was used for years in production cars by the French carmaker
PSA Peugeot-Citroën in the body domain (e.g., for the model) but, as it was not adopted by the
market, it was abandoned in favor of CAN.
13.2.1.3 J1850 Network
heJ[]isanSAEclassBprioritybusthatwasadoptedintheUnitedStatesforcommunications
with nonstringent real-time requirements, such as the control of body electronics or diagnostics. Two
variants of the J are defined: a . kbit/s single-wire version and . kbit/s two-wire version. he
trend in new designs seems to be the replacement of J by CAN or a low-cost network such as
LIN (see Section ...).
13.2.2 Time-Triggered Networks
Among communication networks, as discussed before, one distinguishes TT networks where activ-
ities are driven by the progress of time and event-triggered once where activities are driven by the
occurrence of events. Both types of communication have advantages but one considers that, in gen-
eral, dependability is much easier to ensure using a TT bus (refer, for instance, to Ref. [] for a
discussion on this topic). This explains that, currently, only TT communication systems are being
considered for use in X-by-Wire applications. In this category, multiaccess protocols based on TDMA
are particularly well suited; they provide deterministic access to the medium (the order of the trans-
missions is defined statically at the design time), and thus bounded response times. Moreover, their
