A convoy of driverless U.S. Army trucks uses Lockheed Martin AMAS technology to move along a South Carolina roadway at 40 mph.
OpenECU for Vehicle Supervisory Control in Autonomous Vehicles; Controlling Engine, Braking and Transmission
To apply a common autonomous-operation control appliqué system to a vehicle fleet consisting of varied vehicle types from different manufacturers and have all involved vehicles perform in a similar manner, especially in convoy.
Pi Innovo worked with Lockheed Martin Autonomous Mobility Appliqué System (AMAS), where the complete system consisted of Lockheed Martin’s Overall Vehicle Controller (OVC) plus a vehicle supervisory controller (Pi Innovo’s OpenECU M460) applied to a range of currently-fielded military transport vehicles. At the outset of the project, not all the vehicle types had been identified. As the project progressed, the required list of supported vehicles was expected to grow, but exact configurations of each vehicle were not known in advance.
In order to meet timing requirements and provide a robust, high quality vehicle control system, software architecture decisions were crucial. Software re-use is a common and valuable method for creating robust and efficient systems; this was considered a priority from the outset of the project, particularly since not all vehicle configurations were known in advance.
The hardware for actuating the vehicle consisted of an off-the-shelf Pi Innovo OpenECU M460 rapid controls prototyping unit to control the main vehicle safety-critical systems; brakes, accelerator, compression brake, and transmission via CAN and discrete device controls., The companion Actia Model 2944 Power Management Module (PMM) was used for body control functions such as lighting, turn signals, windshield wipers, etc. Both modules have a successful history of operating in rugged military ground vehicle applications and have a sophisticated user-interface for control configuration.
Using the M460 hardware and OpenECU model based controls Developer Platform SIM-API, Pi Innovo developed the application software that sits between the Lockheed Martin OVC and the vehicle, enabling varied I/O configurations to be created simply and efficiently utilizing a common control core and driving the various vehicle systems in the same manner regardless of physical implementation on the vehicles.
The OpenECU hardware is designed to allow for various levels of customization to provide optional inputs and outputs. In this example, while onsite at Lockheed Martin’s Colorado facility, Pi Innovo engineering was informed of the need for changes of the vehicle electrical system. In a 24 hour period, Pi Innovo’s M460 controller was modified (via on-board component replacements) from its base specification and returned to the vehicle test fleet. The ability to quickly customize the existing proven hardware allowed the development team to continue testing, preserving the project timeline in addition to saving time and money.
Results and Impact
The OpenECU architecture allowed the core of vehicle systems control functionality to be common across all vehicle variants. The OVC made high level navigational decisions and issues commands to the OpenECU M460. The M460 abstracted the OVC from the vehicle and gave true “Appliqué” capability. A mixed fleet utilizing the same command and control methods could be “driven” by the OVC without it knowing which type of vehicle it is driving.
The OpenECU SIM-API platform software provided a flexible controls development environment, allowing the various configurations to be assembled, modified, expanded and verified quickly. The Actia AMPS tool allowed for rapid configuration development for the various body control needs of the vehicle types. Combined, the two controllers were able to accommodate all the variations encountered during the development of the test fleet.
To provide this flexibility, the M460 software was configured to match the target vehicle I/O specification and calibrated to match the performance requirements of the vehicle systems being controlled. The PMM outputs were configured using the AMPS tool, with overall control being provided by the M460 over J1939 CAN.
Systems Architecture Design
Embedded Control Software Development
Rapid Control Prototyping
Control Hardware Design
System Testing and Validation
February 2013-June 2014
For a more in-depth look at AMAS capabilities, click here.