Memory Economy for Electronic Control Units: Compression of Conventional Look-up Tables

The operation of modern technical systems is often controlled by electronic units. In the automotive sector lattice-like look-up tables are a common representation of non-linear multi-dimensional parameter functions of actuators. Regardless of the strongly restricted memory capacity of automotive electronic control units, manufacturers need to introduce additional functionalities to meet the new specifications. With an increasing number of adjustable actuators the number of parameters and thus the memory requirement within the control unit grows enormously. This work introduces a method based on evolutionary and classical optimization algorithms which reduces the memory requirement for 2-dimensional look-up tables. Key-Words: Automotive Electronic Control Unit, Memory Capacity, Look-up Table, nonlinear Optimization, Genetic Algorithm, Evolution Strategy.