© Vetronics Research Centre 1

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set

Dr Panagiotis Oikonomidis

Vetronics Research Centre

University of Brighton

15th international CAN Conference

Vienna

28th October 2015

© Vetronics Research Centre 2

Vetronics Research Centre

The VRC is the only academic centre of excellence in the UK conductingresearch and training in the subject area of Vehicle Electronics(Vetronics) sponsored by the UK MOD and supported by the DefenceScience and Technology Laboratory (Dstl) and Defence Equipment andSupport (DE&S).

Research

We focus on specialised and targeted research to investigate newtechnologies and methodologies that can be applied in the immediateand near future.

Training

The VRC conducts training courses on current and upcoming Vetronicstechnologies and architectures to disseminate our knowledge andexperience within the Vetronics field.

© Vetronics Research Centre 3

Contents

• Sort introduction of CAN FD

• Response time for a CAN message

• Representation of the simulation model

• SAE benchmark based message set

• Performance analysis

• Worst-case delays

• Average delays

• Bus utilisation

© Vetronics Research Centre 4

Introduction

Why is CAN FD necessary

• Increasing amount of electronic control units (ECUs) andmessage traffic

• A rising demand for higher bandwidth and low messageresponse times

CAN FD protocol provides

• Improved bandwidth by using higher bit-rates to transmitthe payload

• Improved payload by increasing the payload size

CAN FD uses the existing CAN environment without requiringsignificant changes in hardware and software structure for asmooth transition

© Vetronics Research Centre 5

Response time for a CAN message

𝑹𝒎 = 𝑱𝒎 +𝑾𝒎 + 𝑪𝒎Jm is the queuing jitter, the latest queuing time of message mm on the host CPU

Wm includes the queuing delay due to the higher priority messages and the blockingtime due to a lower priority message

Cm is the time taken to transmit message mm on the bus

Message transmission time (Cm) with worst-case bit stuffing for a standard CANframe:

𝑪𝒎 = 𝟒𝟕 + 𝟖𝒔𝒎 +𝟑𝟒+𝟖𝒔𝒎

𝟒𝝉𝒂

sm is the payload size in bytes

τ𝑎 is the bus bit-time

Message transmission time (Cm) with worst-case bit stuffing for a standard CAN FDframe:

𝑪𝒎 = 𝟐𝟗 +𝒔𝒂𝒓𝒃

𝟒𝝉𝒂+ 𝟐𝟖 + 𝟖𝒔𝒎 +

𝟖𝒔𝒎

𝟒+ 𝑪𝑹𝑪 𝝉𝒅

sarb is the number of bits subject to bit-stuffing in the first arbitration-phase

τd is the data-phase bit-time

© Vetronics Research Centre 6

Simplified representation of the simulation model

• The models have been simulated at 1 Mbps, 500 kbps, 250 kbps, and 125kbps transmission bit-rates.

• 5 Mbps is used as the CAN FD data-phase transmission bit-rate

© Vetronics Research Centre 7

Modelling and Simulation

Simulation model developed with SimEvents Toolbox of Matlab Simulink

© Vetronics Research Centre 8

Simulation Components of a Node

© Vetronics Research Centre 9

Simulation Components of CAN Bus

© Vetronics Research Centre 10

SAE benchmark based message set

• Messages are placed in priority order from the highest (M1) to the lowest (M17)

• T indicates the message period, while D indicates the deadline in milliseconds

MessageSize

(bytes)T (ms) D (ms)

M1 1 50.0 5.0

M2 2 5.0 5.0

M3 1 5.0 5.0

M4 2 5.0 5.0

M5 1 5.0 5.0

M6 2 5.0 5.0

M7 6 10.0 10.0

M8 1 10.0 10.0

M9 2 10.0 10.0

M10 3 10.0 10.0

M11 1 50.0 20.0

M12 4 100.0 100.0

M13 1 100.0 100.0

M14 1 100.0 100.0

M15 3 1000.0 1000.0

M16 1 1000.0 1000.0

M17 1 1000.0 1000.0

© Vetronics Research Centre 11

Performance analysis

• Worst-case message delays

• Average message delays

• Bus utilisation

© Vetronics Research Centre 12

Worst-case delays

Worst-case delays at 1 Mbps Worst-case delays at 500 kbps

Worst-case delays at 250 kbps Worst-case delays at 125 kbps

© Vetronics Research Centre 13

Worst-case message delay ratios

Delay performance ratio = CAN message delay / CAN FD message delay

• At 1 Mbps, compared to CAN, the worst-case CAN FD message delay is 1.78 timessmaller on average.

• At 500 kbps, CAN FD has half the CAN worst-case delay on average.

• The ratio is 2.16 and 3.28 at 250 kbps and 125 kbps bit rates, respectively.

CAN / CAN FD 1 Mbps 500 kbps 250 kbps 125 kbps

Average 1.78 2.00 2.16 3.28

Minimum 1.66 1.88 1.99 2.05

Maximum 1.98 2.15 2.39 6.00

© Vetronics Research Centre 14

Average delays

Average delays at 1Mbps Average delays at 500 kbps

Average delays at 250 kbps Average delays at 125 kbps

© Vetronics Research Centre 15

Average message delay ratios

Average ratio from 1 Mbps to 125 kbps ranges from 1.67 to 4.16, up to4.16 times smaller mean message delays are achieved on average withCAN FD.

CAN / CAN FD 1 Mbps 500 kbps 250 kbps 125 kbps

Average 1.67 1.94 2.31 4.16

Minimum 1.53 1.73 1.92 2.16

Maximum 2.17 2.62 3.03 6.00

© Vetronics Research Centre 16

Bus utilisation

The “CAN reference” and “CAN” graphs have almost the samecharacteristics. The “CAN FD” graph shows that the same message set istransmitted with less bus utilisation with CAN FD.

Bus utilisation values from 1 Mbps to 125 kbps range from 10.3 to 82.5for CAN, and from 6.7 to 43.8 for CAN FD, respectively.

© Vetronics Research Centre 17

Conclusion

CAN FD protocol provides considerable performance improvements in:

• message transmission speed

• bus utilisation

to meet the requirements of real-time control systems.

© Vetronics Research Centre 18

Any questions?