Design Methods to Optimize the Performance of Controller Area Networks 2012-01-0194
This literature is in the field of communication networks where different Electronic Control Units (ECUs) communicate with each other over Controller Area Network (CAN) protocol. Typically these types of CAN networks are widely used in automotive vehicles, plant automations, etc. This proposed method is applicable in all such applications where controller area network is used as backbone electrical architecture.
This literature proposes a new method of CAN signal packing into CAN frames so that network bus-load is minimized so that more number of CAN signals can be packed and more number of ECUs can be accommodated within a CAN network. The proposed method also ensures that the age of each CAN signal is minimized and all CAN signals reach the intended receiving ECUs within their maximum allowed age. Typically network designers are forced to design and develop multiple sub-networks and network gateways to get rid of network bus-load. As the proposed method intends to minimize network bus-load, the requirement of gateways just to reduce bus-load will be avoided.
The proposed method also proposes method for assignment of CAN Identifiers in each CAN frame by introducing different sub-fields within the CAN Identifiers. This will help to reduce the delay in the CAN frames due to arbitration loss in the network. The proposed methods will minimize the ECU loads network due to CAN frame reception for all ECUs in the CAN by minimizing the number of CAN frames to be received by each ECU in the network and also by enabling hardware filtering of CAN frames by receiving ECUs due to different sub-fields within the CAN Identifier.
This paper also proposes method to distribute different CAN frames with time to avoid burst transmission of CAN frames by different ECUs in the CAN network and hence reducing the possibility of frame loss in the network.