Lookup Table Estimation
Estimators based on lookup tables are commonly used in engine control units (ECUs), battery systems, and other embedded applications to represent physical behavior that is difficult or costly to measure directly in production. Lookup table reduced-order-models (ROM) are also widely used in system‑level simulations when a detailed physical model is impractical in terms of execution speed and memory footprint.
In model-based calibration (MBC) workflows, the term feature is often used to describe a controller feature that represents a distinct vehicle behavior or function that is characterized by one or more lookup tables. These tables encode calibration data that maps operating conditions to system response.
Model-Based Calibration Toolbox™ uses a Simulink® model of the lookup table‑based system, built from a defined subset of Simulink blocks. The calibration process compares the lookup table model to a reference data set, which can come from measured system data, simulation of a detailed physical model, or a statistical response surface fitted to one or more data sets. This process minimizes the difference between the lookup table model and the reference data set.
Start with the Process for Calibrating and Filling Feature Tables.
Apps
| MBC Optimization | Generate optimal lookup tables for model-based calibration |
Functions
evaluateFeature | Evaluate feature in CAGE project with data set (Since R2024b) |
characterizeBattery | Characterize equivalent circuit battery in CAGE project (Since R2024b) |
exportCalibration | Export calibration items from CAGE project (Since R2024b) |
Topics
Feature Calibration Setup
- Process for Calibrating and Filling Feature Tables
Calibrate and fill feature lookup tables. - Set Up Feature and Strategy
Add a feature and import or construct a strategy. - Optimize Calibrations with Transient Controllers
Use the transient feature fill to optimize calibrations with transient controllers.
Lookup Tables
- Initialize Feature and Calibration Tables
Initialize a feature to set the values of the normalizers over the variable ranges. - Fit Data to N-D Lookup Table Using fitlookupn
Fit data to n-d lookup table using fitlookupn. - Fill Feature Lookup Tables
Initialize, fill, extrapolate, and optimize your table values. - Fill a Single Lookup Table
Fill a lookup table with measured data or from a model evaluated at the table breakpoints.
Optimization and Validation
- Calibrate a Feature for an Estimator Subsystem in an ECU
Calibrate an ECU subsystem that estimates engine torque as a function of speed, load, AFR, and spark angle.
- STEP 1: Set Up Feature and Torque Strategy
- STEP 2: Import Two-Stage Models
- STEP 3: Align Model and Feature Inputs
- STEP 4: Fill Feature and Calibration Tables
- STEP 5: Validate Feature and Export Lookup Tables
- Validate Feature Fill
Use validation data to validate the feature fill.
Battery Model Calibration
- Calibrate Powertrain Blockset and Simscape Battery Blocks
Adjust lookup tables in Simscape™ Battery™ and Powertrain Blockset™ blocks with experimental data. (Since R2023a) - Characterize Battery Equivalent Circuit Block (Code)
Characterize the Battery Equivalent Circuit block using thecharacterizeBatterymethod. - Characterize Powertrain Blockset and Simscape Battery Blocks (App)
Characterize Simulink or Simscape battery block parameters using the MBC Optimization app.
- Characterize an Equivalent Circuit Battery
Use CAGE to characterize an equivalent circuit battery.
