Proceedings of the IEEE International Conference on Automation and Logistics Shenyang, China August 2009
Temporal Characteristic Constraints between SP and PG of SAE J1939 in Passenger Car Information Integrated Control System Jian HU*, Gangyan LI* and Yeqiong SONG** * School of Mechanical and Electronic Engineering
** LORIA - Equipe TRIO
Wuhan University of Technology
615, Rue Du Jardin Botanique, 54600
122 Luoshi Road, 430070, Wuhan, Hubei Province, China
Villers Les Nancy, France
[email protected],
[email protected]
[email protected]
Abstract - As a higher layer protocol based on CAN (Controller Area Network), SAE J1939 has specified SP (Suspect Parameter) and PG (Parameter Group) which respectively represent signal exchanged between ECUs (Electronic Control Units) and message containing a set of SPs. There is temporal characteristic constraint, such as period and deadline, between SP and PG in SAE J1939-based passenger car information integrated control system. According to the periods of SPs, it is easy to deduce that the period of PG is the minimal period of SPs composing this PG. Considering the possible offset between the production time of data of SP and the first production time of PG containing this data, the calculation formula of deadline of PG is deduced by applying the relationship between the production times of two periodic SPs. Vice versa, the given temporal characteristic of PG will also constrain the period and deadline of SPs composing this PG. According to two requirements, (1) deadline of PG equals to the minimal deadline of SPs composing this PG and (2) temporal characteristic of PG containing these SPs has been specified in advance, the temporal characteristic constraints of SPs are put forward.
and MOST (Media Oriented Systems Transport) [12] which have different characteristics and are applied in different automotive domains. Nowadays, CAN is widely applied in different automotive control systems thanks to its easy use, flexibility and robustness. As a higher layer protocol based on CAN, SAE J1939 [13] has been developed by SAE (Society of Automotive Engineers) and allows ECUs to communicate with each other by providing a standard architecture. In SAE J1939-based passenger car information integrated control system, SP (Suspect Parameter) [14] denotes the parameter related to ECUs and transmitted through in-vehicle network. The network transmission of SP is implemented by PG (Parameter Group) [14], i.e. SAE J1939 message which is the information unit exchanged in data link layer. During transmission process, one SAE J1939 message contains several data of SPs in the interest of optimizing network performance. In order to provide consistent information in V mode-based system development process, SAE J1939 communication database have been designed to implement management functions of ECUs, PGs and SPs [2]. In [15], SAE J1939 simulation model also has been established to evaluate network performance. Period and deadline are primary parameters to describe temporal characteristics of SP and PG. In the design process of SAE J1939-based passenger car information integrated control system, some periods of PGs have been specified in SAE J1939 application layer. The deadlines of these PGs are known if it is assumed that the deadline of PG equals to the period of PG in SAE J1939. However, the temporal characteristics of the remainder PGs are not specified but designed by system engineer or control engineer. In essence, the temporal characteristic of SP depends on sampling period of sensor, control algorithm and experience of engineer etc., the period and deadline of PG depend on the temporal characteristics of SPs composing this PG. But there is no research on the temporal characteristic constraint between SP and PG of SAE J1939 currently. It doesn’t consider the effect on temporal characteristic of PG deriving from that of SPs. In the same way it also omits the temporal characteristic constraint of SP when the period and deadline of PG have been provided in advance. In this paper, we mainly focus on the temporal characteristic constraint between SP and PG of SAE J1939 in passenger car information integrated control system. The
Index Terms –Automotive Electronic; Passenger Car; SAE J1939; Real Time; Temporal Characteristic
I. INTRODUCTION Passenger car is the main traffic tool in public transportation and equipped with many ECUs (Electronic Control Units) to fulfil with the requirements of adding functions and enhancing performance. The implementation of complicated control function requires the use of real-time and reliable information exchange between different ECUs, sensors and actuators. But it is hard to realize information exchange with the traditional point-to-point links which would induce drawbacks such as the increase of weight, cost and reliability problem etc. [1]. Based on in-vehicle network technology, passenger car information integrated control system can share information and implement real-time and correlative control between ECUs [2]. At the beginning of 1980s the engineers of the automotive manufacturers assessed the existing field bus systems for their use in vehicles. With the development of in-vehicle network technology, the typical in-vehicle network protocols are LIN (Local Interconnect Network) [3], CAN (Controller Area Network) [4]-[6], TTCAN (Time Triggered CAN) [7]-[9], TTP/C (Time Triggered Protocol/Class C) [10], FlexRay [11],
978-1-4244-4795-4/09/$25.00 © 2009 IEEE
359
remainder contents of this paper are organized as follows: In section II, SAE J1939 protocol and its application layer, especially SP and PG, are introduced. By discussing the temporal characteristics of SP and PG with an instance, in section III three problems which should be resolved about the temporal characteristic constraint are presented. In section IV, the calculation formula of deadline of PG according to the temporal characteristics of SPs composing this PG is deduced. Because the temporal characteristic of PG also constrains that of SPs composing this PG, in section V, the temporal characteristic constraints of SPs under two conditions, (1) the deadline of PG equals to the minimal deadline of SPs composing this PG and (2) temporal characteristic of PG containing these SPs has been specified in advance, are put forward. Finally, the work of this paper and the future work are concluded and presented respectively in section VI.
PG is unconcerned with the source address of ECU, it can be transmitted from any source address. One PGN (Parameter Group Number) identifies one PG uniquely [14]. III. TEMPORAL CHARACTERISTICS PROBLEM OF PG OF SAE J1939 Period and deadline are primary parameters to describe temporal characteristics of SP and PG in SAE J1939-based passenger car information integrated control system. Some periods of PGs have been specified in SAE J1939 application layer. The deadlines of these PGs are known if it is assumed that the deadline of PG equals to the period of PG. However, Similar to LC of Table I, the temporal characteristics of remainder PGs, especially for aperiodic PGs which should be transformed to periodic PGs, are not specified but designed by system engineer or control engineer.
II. SAE J1939 AND ITS APPLICATION LAYER
Table I Definition of Lighting Command -LC Lighting Command - LC On change of lamp on/ off state. Transmission Rate: 8 bytes Data Length: 0 Data Page: 254 PDU Format: 65 PDU Specific: 3 Default Priority: PGN 65089(FE41) POS Length Parameter Name 1.1 2 bits Running Light Command Alternate Beam Head Light 1.3 2 bits Command 1.5 2 bits Low Beam Head Light Command 1.7 2 bits High Beam Head Light Command … … …
Because CAN specification just defines physical layer and data link layer based on ISO/OSI Reference Model, it is convenient for automotive manufacturers to specify their own CAN application layers. Some examples of CAN-based application layer protocols are CANopen, DeviceNet and SAE J1939 which are very special and related to specific application fields. As a higher layer protocol based on CAN, SAE J1939 has been developed by the SAE Truck & Bus Control and Communications Network Subcommittee of the Truck & Bus Electrical & Electronic Committee. The purpose of SAE J1939 is to allow ECUs to communicate with each other by providing a standard architecture. SAE J1939 provides a complete network definition using the 29 bits identifier (CAN extended frame) defined within CAN protocol and specifies the formats, such as length, limit and type of data, of common signals. As the two basic parameters of SAE J1939, SP and PG are defined in details in SAE J1939 application layer.
SPN 2403 2351 2349 2347 …
In the design process of SAE J1939-based passenger car information integrated control system, the period and deadline of SP depends on sampling period of sensor, control algorithm and experience of engineer etc., at the same time the period and deadline of PG depends on the temporal characteristics of SPs in essence. It is assumed that PG of SAE J1939, symbolized with pg k
A. SP and SPN In SAE J1939 application layer, SP is used to identify a particular parameter, element or component associated with ECUs. The data of SP transmitted through network is not always equal to the actual value of parameter. In order to define the actual value of parameter, SAE J1939 application layer specifies SLOT [14] for each SP where S, L, O and T respectively denote the scaling between the values of SP and actual parameter, the limit of actual value of parameter, the offset, i.e. initial value, of the actual value of parameter and the transfer function of parameter. One SPN (Suspect Parameter Number) identifies one SP uniquely. The calculation formula between the values of SP and the actual parameter is described as follows: Actual Value of Parameter = Scaling × Value of SP + Offset (1)
( k ∈ Ζ + ), consists of SP set { sp1k , sp2k , …, spnk } ( n ∈ Ζ + ). •
The periods of pg k and spik ( i ∈ Ζ + ) are Tk ( Tk ∈ Ζ + ) and Ti •
( Ti ∈ Ζ + ) respectively. Every SP of { sp1k , sp2k , …, spnk } has k the same first production time. sp min is the SP with the •
minimal period of { sp1k , sp2k , …, spnk }, i.e. T min = min •
{ Ti | spik ∈ { sp1k , sp2k , …, spnk }}. It is also assumed that the first production time of pg k is synchronized with the first k production time of sp min .
B. PG and PGN SAE J1939 application layer groups several SPs into one message transmitting through network so as to use the 8 bytes data field efficiently. PG defines SAE J1939 message containing SPs in details. The description of PG should include the characteristics such as transmission rate, data length, priority and definition of position of SPs etc. Because
It easily deduces that for ∀ spik ∈ { sp1k , sp2k , …, spnk }, if
pg k is transmitted with the period: •
Tk = T min
360
(2)
•
deadline of SPs composing this PG, what temporal characteristic constraints should be imposed on these SPs? (3) Considering the period and deadline of PG having been specified in advance, what temporal characteristic constraints should be imposed on SPs composing this PG? The solution of the first problem is the precondition to resolve the other two problems. The focus of temporal characteristic of PG is to find the deadline calculation method of PG.
the data of spik produced at ki ∗ Ti (k ∈ Ν ) can be i •
transmitted before (ki + 1) ∗ Ti . But the deadline of pg k is not the minimal deadline of { sp1k , sp2k , …, spnk }. It is considered the example shown on Fig.1 with sp1 and •
•
sp2 respectively having periods T1 = 5 , T2 = 8 and deadlines •
•
D1 = 5 , D2 = 8 . sp1 and sp2 are synchronized at time 0. According to formula (2), the period of PG containing sp1 and •
•
IV. DEADLINE CALCULATION METHOD OF PG OF SAE J1939 A. Deadline Calculation of PG of SAE J1939 It is assumed that PG of SAE J1939, symbolized with pg k
•
sp2 is equal to min(T1 , T2 ) = T1 = 5 . As shown in Fig.1, the data of sp2 produced at time 8 is actually transmitted at time 10 and the offset between the production time and transmission time is 10 − 8 = 2 . Because the invalidation time of this data is 16 , thus the deadline of PG must be less than or equal to 16 − 10 = 6 in order to respect the temporal constraint of sp2 . The data of sp2 produced at time 16 is actually transmitted at time 20 and the offset between the production time and transmission time is 20 − 16 = 4 . Because the invalidation time of this data is 24 , thus the deadline of PG must be less than or equal to 24 − 20 = 4 . It shows that this
( k ∈ Ζ + ), consists of SP set { sp1k , sp2k , …, spnk } ( n ∈ Ζ + ). •
The period and deadline of spik ( i ∈ Ζ + ) is Ti ( Ti ∈ Ζ + ) and •
•
Di ( Di ∈ Ζ + ) respectively. Every SP of { sp1k , sp2k , …, spnk } k has the same first production time. sp min is the SP with the •
minimal period of { sp1k , sp2k , …, spnk }, i.e. T min = min •
k { Ti | sp min ∈ { sp1k , sp2k , …, spnk }}. The period and deadline of
value is less than the deadline of sp1 , D1 = 5 . This example
pg k is Tk ( Tk ∈ Ζ + ) and Dk ( Dk ∈ Ζ + ) respectively. It is also assumed that the first production time of pg k is synchronized
illustrates that the deadline of pg k is not the minimal deadline
k with the first production time of sp min .
•
•
of SP set { sp1k , sp2k , …, spnk } because there is possible offset [16] between the production time of the data of SP and the first transmission time of the PG containing this data. 8
0
sp 2
16
24
As shown in Fig.2, Offset (Tk , Ti ) is defined as the time offset between the production time of the data of spik and the first
production
""
time
pg k
of
containing
this
data;
•
Offsetmax (Tk , Ti ) is defined as the maximal value of
t
•
Offset (Tk , Ti ) . 16 − 10 = 6
24 − 20 = 4
5
0
10
15
20
t
•
sp1 and sp2 respectively having periods T1 = 5 , T2 = 8
•
and deadlines
""
t
""
25 •
Fig.1 Two SPs
""
spik
sp1
•
""
D1 = 5 , D2 = 8 . The period of PG containing sp1 and sp2
""
pg k •
•
is equal to
Offset (Tk， Ti )
T1 = 5 .
•
Offset (Tk， Ti )
•
Offset max (Tk， Ti )
t
•
Offset (Tk， Ti )
•
Fig.2
The calculation problem of the deadline of pg k can be
Offset (Tk , Ti ) : time offset between the production time of the data of
spik and the first pro mduction time of pg k containing this data;
described as follows: for ∀ spik ∈ { sp1k , sp2k , …, spnk }, find
•
•
out the value of Dk which ensures that the response time of
Offsetmax (Tk , Ti ) : the maximal value of Offset (Tk , Ti ) .
spik satisfies with the deadline constraint of spik if the response time of pg k satisfies with the deadline constraint of pg k . Therefore, three problems that should be resolved about the temporal characteristic constraints between SP and PG of SAE J1939 are presented: (1) How to calculate the period and deadline of PG according to the temporal characteristics of SPs composing this PG? (2) Aiming at the deadline of PG being equal to the minimal
Because it exists possible time offset between the production time of the data of spik and the first production time of pg k containing this data, the deadline of pg k is •
defined as the minimal value of the difference between Di and •
Offsetmax (Tk，Ti ) , i.e.: •
•
Dk = min{Di − Offset max (Tk， Ti ) | spik ∈ {sp1k，sp 2k， "，sp nk }} (3)
361
Thus the key of calculating the deadline of pg k is to
•
=(
•
calculate Offset max (Tk， Ti ) . •
•
•
ti = ki ∗ Ti (ki ∈ Ν )
•
tk − ti = k min ∗ T min − ki ∗ Ti ≥ 0 According to the definition of time offset:
0 ≤ k min ∗ T min − ki ∗ Ti < T min Applying Ref [17], the production times of any two datum of two periodic SPs have the relationship as: •
p1 − p2 = q ∗ gcd(T1 ，T2 ) , i.e.: •
k1 ∗ T1 − k2 ∗ T2 = q ∗ gcd(T1 ，T2 ) •
•
•
•
Setting T min = T1 , Ti = T2 •
•
•
•
Thus k min ∗ T min − ki ∗ Ti = q ∗ gcd(T min ，Ti ) ( q ∈ Ν ) Thus the key of calculating the maximal value of •
•
k min ∗ T min − ki ∗ Ti is to calculate the maximal value of •
•
q ∗ gcd(T min ，Ti ) . •
Because gcd(T min ，Ti ) is a constant, Thus the key of calculating the maximal value of •
k min ∗ T min − ki ∗ Ti is to calculate the maximal value of q . The maximal value of q is deduced as follows: •
•
•
Because 0 ≤ k min ∗ T min − ki ∗ Ti < T min , i.e.: •
•
•
0 ≤ q ∗ gcd(T min ，Ti ) < T min •
Thus q <
T min
•
•
•
gcd(T min ，Ti )
•
•
, and because gcd(T min ，Ti ) | T min ,
•
Thus max(q) =
T min •
•
gcd(T min ，Ti )
•
•
•
•
Table II Period and deadline of some SPs of LC Period Parameter Name SPN (ms) Running Light Command 2403 500 Alternate Beam Head Light Command 2351 80 Low Beam Head Light Command 2349 80 High Beam Head Light Command 2347 80 Tractor Front Fog Lights Command 2387 100 Right Turn Signal Lights Command 2369 50 Left Turn Signal Lights Command 2367 50 Back Up Light and Alarm Horn 2391 50 Command Centre Stop Light Command 2375 100 Tractor Marker Light Command 2377 500 Rear Fog Light Command 2389 100 Tractor Underside Mounted Work Lights 2357 500 Command Lighting Data Request Command 2393 500
•
•
•
B. Analysis of the Temporal Characteristic of LC There are many light control signals in passenger car information integrated control system. In SAE J1939 application layer, the PG, Lighting Command-LC, contains almost all the SPs of light commands especially for truck. Because of different importance, there are different sampling period and real time request for each SP of light commands. Therefore, the periods and deadlines of these SPs are different. As shown in Table II, the periods and deadlines of some SPs of light commands used in WG6100ENH city bus are defined. It is assumed that the deadline of each SP of light commands equal to its period. Because there are differences of light names between passenger car and truck, we adopt the SAE J1939 light names which don’t affect on the result of the temporal characteristic of PG.
•
•
•
•
•
•
•
•
Dk = min{Ti − T min + gcd(T min ， Ti ) | spik ∈ {sp1k，sp2k， "，spnk }} (5)
tk − ti < T min , i.e.:
•
•
If Di = Ti ,
•
•
•
Dk = min{Di − T min + gcd(T min ， Ti ) | spik ∈ {sp1k，sp 2k， "，spnk }} (4)
•
•
•
= min{Di − T min + gcd(T min ， Ti ) | spik ∈ {sp1k，sp2k， "，spnk }} Therefore, the deadline of pg k :
Thus t k = k min ∗ T min (kmin ∈ Ν ) , •
•
•
•
•
•
Dk = min{ Di − Offset max (Tk， Ti )} | spik ∈ {sp1k，sp2k， "，spnk } }
According to formula (2), Tk = T min ,
•
gcd(T min ，Ti )
•
− 1) ∗ gcd(T min ，Ti )
•
•
data. Thus Offset (Tk，Ti ) equals to tk − ti .
•
•
= T min − gcd(T min ， Ti ) According to formula (3),
spik , t k is the first production time of pg k containing this
•
•
•
•
It is assumed that ti is the production time of the data of •
T min
−1
•
Deadline (ms) 500 80 80 80 100 50 50 50 100 500 100 500 500
•
T min = min { Ti | sik ∈ { s1k , s 2k , …, s nk }} = 50(ms) Applying formula (2),
Thus:
•
•
TLC = T min = 50(ms) Applying formula (4),
Offsetmax (Tk，Ti ) •
•
= max(k min ∗ T min − k i ∗ Ti ) •
•
= max(q ∗ gcd(T min ，Ti )) •
•
•
•
DLC = min{Di − T min + gcd(T min ，Ti ) | spik ∈ {sp1k， sp2k，"， spnk }}
•
= 40(ms) According to the calculation results, it is shown that the deadline of LC isn’t always equal to the minimal deadline of SPs composing it. It is essential to obtain the deadline of PG
•
= max(q) ∗ gcd(T min ，Ti )
362
•
with formula (4) because it will affect on the analysis of system schdulability and the selection of nework transmission rate if we configure the deadline of PG as the the minimal deadline of SPs composing this PG.
•
•
•
•
deadline of pg k are Tk ( Tk ∈ Ζ + ) and Dk ( Dk ∈ Ζ + ) respectively. It is also assumed that the first production time of pg k is synchronized with the first production time of k spmin( . T) According to formula (4), the difference between the
•
and Offsetmax (Tk，T min( D ) ) is : •
•
the
of spik
deadline
and
Thus in order to ensure Tk , it is required: •
Offsetmax (Tk，Ti ) is : •
Ti ≥ Tk According to formula (4),
•
Di − T min + gcd(T min ，Ti )
•
•
•
•
•
•
•
•
•
•
•
•
•
If Ti = Di , it can be considered that sp
k min( T )
•
and sp
•
•
(9)
•
If Ti = Di and Dk = Tk , formula (9) can be transformed to:
(6) k min( D )
•
Di ≥ Dk + Tk − gcd(Tk，Ti )
•
Di ≥ D min + gcd(T min ，T min( D ) ) − gcd(T min ，Ti )
•
Di − Tk + gcd(Tk，Ti ) ≥ Dk , i.e.:
i.e. the temporal characteristics of spik should satisfy with: •
•
In order to ensure Dk , it is required
•
Di − T min + gcd(T min ，Ti ) ≥ D min − T min + gcd(T min ，T min( D ) ) •
•
(8)
Dk = min{Di − T min + gcd(T min ，Ti )}
•
If it is required Dk = D min , the upper two formulas must satisfy with: •
•
Applying formula (2), the period of pg k : Tk = T min
•
•
•
The constraint conditions of Ti and Di will be deduced as follows.
•
D min − T min + gcd(T min ，T min( D ) )
•
•
B. Temporal Characteristic Constraints of SPs under the Condition that Temporal Characteristic of PG Containing these SPs Has Been Specified in Advance The possibility to add new functions means add new ECUs in passenger car information integrated control system and adding more ECUs implies that more SPs will be exchanged through in-vehicle network. The temporal characteristic of PG specified in advance will also constrain the periods and deadlines of SPs composing this PG. At the same time, in order to decrease the influence on real time performance of passenger car information integrated control system, it is hoped that the period and deadline of PG are not changed when a new SP is inserted into this PG. The inserted SP should satisfy with special temporal characteristic constraint for achieving this purpose. Therefore, these two problems can be summarized to the problem of temporal characteristic constraints of SPs aiming at the temporal characteristic of PG containing these SPs having been specified in advance. The assumption of this problem is the same as Section IV.
•
between
•
•
•
difference
•
= 100 − gcd(50，Ti ) (ms)
D min = min{ Di | spik ∈ { sp1k , sp2k , …, spnk }}.The period and
The
•
Ti ≥ 2 × T min − gcd(T min ，Ti )
where T min = min{ Ti | spik ∈ { sp1k , sp2k , …, spnk }},
•
•
Then the temporal characteristic of spik should satisfy with:
k and sp min( are T min , D min(T ) and T min( D ) , D min respectively D)
•
•
•
k sp2k , …, spnk } respectively. The period and deadline of spmin( T)
•
•
DLC = D min = 50 (ms)
SPs with the minimal period and minimal deadline of { sp1k ,
deadline of sp
•
•
•
k k and sp min( are the has the same first production time. spmin( T) D)
k min( D )
•
Ti ≥ 2 × T min − gcd(T min ，Ti ) (7) Regarding the LC of Table III as instance, if it requires:
•
•
•
characteristics of spik should satisfy with:
Di ( Di ∈ Ζ + ) respectively. Every SP of { sp1k , sp2k , …, spnk }
•
•
i.e., if Ti = Di and it is required Dk = D min , the temporal
•
•
•
= 2 × T min − gcd(T min ，Ti )
The period and deadline of spik ( i ∈ Ζ + ) is Ti ( Ti ∈ Ζ + ) and
•
•
•
( k ∈ Ζ + ), consists of SP set { sp1k , sp2k , …, spnk } ( n ∈ Ζ + ).
•
•
Ti ≥ T min + gcd(T min ，T min ) − gcd(T min ，Ti )
V. TEMPORAL CHARACTERISTIC CONSTRAINTS OF SPS DERIVING FROM THE TEMPORAL CHARACTERISTIC OF PG A. Temporal Characteristic Constraints of SPs under the Condition that Deadline of PG Equals to the Minimal Deadline of SPs Composing this PG In the design of SAE J1939-based passenger car information integrated control system, for convenience of system schedulability analysis, it is required that the deadline of PG equals to the minimal deadline of SPs composing this PG especially when the deadline of SP equals to the period of SP. Aiming at this purpose, the temporal characteristic of SP should satisfy with special constraint condition. It is assumed that PG of SAE J1939, symbolized with pg k
•
•
Therefore, D min = T min( D ) = T min . Formula (6) can be transformed as:
•
•
Ti ≥ Tk + Tk − gcd(Tk，Ti ) , i.e.
are
k . the same SP, denoted as sp min
•
•
Ti ≥ 2 × Tk − gcd(Tk，Ti )
363
(10)
If the constraint conditions of formula (10) is satisfied, the
We would like to give our acknowledgments to Wuhan city, Hubei province, and Wuhan University of Technology, China because the research work reported here has been sponsored by “Wuhan city tackle key science and technology problem project” (200851430478, 200710421118), “Hubei province digital manufacturing key laboratory open foundation” (SZ0615) and “Excellent PhD. Dissertation Sustentation Fund of Wuhan University of Technology”(200704). We also want to thank Mathieu Grenier and Nicolas Navet, LORIA, France, for their helpful advices on this study.
•
period of Ti must accord with formula (8) because •
•
Ti ≥ 2 × Tk − gcd(Tk，Ti ) ≥ 2 × Tk − Tk = Tk according to formula (9). •
•
Therefore, If Ti = Di and Dk = Tk , in order to ensure Tk and
Dk of pg k , it requires: •
•
Ti ≥ 2 × Tk − gcd(Tk，Ti ) Regarding the ETC1 of SAE J1939 [14], if it is required TETC 1 = 10 (ms), in order to ensure TETC 1 , according to formula (8) any period of SPs should satisfy with:
REFERENCES [1]
•
Ti ≥ 10 (ms)
[2]
If it is specified DETC 1 = TETC1 = 10 (ms), in order to ensure
DETC 1 , according to formula (9) any deadline of SPs should satisfy with: [3]
•
•
Di ≥ DETC 1 + TETC1 − gcd(TETC1，Ti )
[4]
•
= 20 − gcd(10，Ti ) (ms) •
[5] [6]
•
If it is assumed Ti = Di , in order to ensure TETC 1 and DETC 1 of pg ETC 1 , according to formula (10) any period of SPs should satisfy with: •
[7]
•
[8]
Ti ≥ 20 − gcd(10，Ti ) (ms)
VI. CONCLUSIONS
[9]
In this paper, we focused on the temporal characteristic constraints between SP and PG of SAE J1939 in passenger car information integrated control system. We presented that the period of PG is the minimal period of SPs composing this PG. Based on the mathematic formula of the difference between the production times of two periodic SPs, we deduced the calculation formula of deadline of PG according to the temporal characteristics of SPs composing this PG. According to the requirements of temporal characteristic of PG, we put forward the temporal characteristic constraints of SPs under two conditions. SAE J1939 utilizes all of 8 bytes data field to group SPs into one PG so as to reduce the waste of data field of message. However, it doesn’t avoid the waste of data field because SAE J1939 is specified especially for truck and many SPs of SAE J1939 are useless for other vehicles. Moreover, some signals of other vehicles, such as passenger car, are not specified in SAE J1939. The further research will focus on optimizing the number of PGs according to special applications. In addition, we are also concerned with the PGs assignment method aiming at optimizing network performance based on the results of this paper.
[10] [11] [12] [13] [14] [15]
[16] [17]
ACKNOWLEDGMENT
364
Miguel A. Domínguez, Perfecto Mariño, Francisco Poza, et al, “Communication Networks for Vehicular Electronic Devices”, in Proceedings of EUROCON 2007 The International Conference on “Computer as a Tool”, Warsaw, 2007, pp.1061-1067. Jian Hu, Gangyan Li, Xiangpeng Yu, et al, “Design and Application of SAE J1939 Communication Database in Passenger car Information Integrated Control System Development”, in Proceedings of the 2007 IEEE International Conference on Mechatronics and Automation (IEEE ICMA 2007), harbin, china, 2007, pp.3429-3434. LIN Consortium, “LIN specification package, revision 2.0”, Online available: http://www.lin~subbus.org/, Feb 2005. ISO, “Road vehicles-low speed serial data communication-part 2: low speed controller area network”, in ISO 11519-2, 1994. Bosch, “CAN Specification 2.0”, in Robert Bosch GmbH, 1991. ISO, “Road vehicles-interchange of digital information-controller area network for high-speed communication”, in ISO 11898, 1994. B. Müller, T. Führer, F. Hartwich, et al, “Fault tolerant TTCAN networks”, in Proceedings 8th International CAN Conference, LasVegas, 2002, pp. 3-9. Thomas Führer, B Müller, W Dieterle, et al, “Time triggered communication on CAN”, in Proceedings 7th International CAN Conference, 2000, pp.1-7. ISO, “Road vehicles-controller area network (CAN)-part 4: timetriggered communication”, in ISO 11898-4, 2000. TTTech Computertechnik GmbH, “Time-triggered protocol TTP/C, high-level specification document, protocol version 1.1”, Online available: http://www.tttech.com, Aug 2007. FlexRay Consortium, “FlexRay communication system, protocol specification, version 2.0”, Online available: http://www.flexray.com, Feb 2007. MOST Cooperation, “MOST specification revision 2.3”, Online available: http://www.mostnet.de, Jan 2008. SAE, “Surface vehicle recommended practice, SAE J1939”, in SAE J1939, Jan 2005. SAE, “Surface vehicle recommended practice, SAE J1939-71”, in SAE J1939, Jan 2005. Jian Hu, Gangyan Li, Xiangpeng Yu, et al, “CAN Modeling and Network Simulation for Passenger car Information Integrated Control System”, in Proceedings of the 2007 IEEE International Conference on Mechatronics and Automation (IEEE ICMA 2007), harbin, china, 2007, pp. 1718-1723. Rishi Saket, Nicolas Navet,”Frame packing algorithms for automotive applications”, in Journal of Embedded Computing, vol.2, 2006, pp. 93102. Jian Hu, Gangyan Li, Yeqiong Song, “Research on Calculation Method of Period and Deadline of Frame in Automotive Electronic and Information Integrated Control System”, in International Journal of Factory Automation, Robotics and Soft Computing, no.1, 2009, pp. 1723.
