AI For EV

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

AI Applications in EV

Image credits: Digikey


AI Landscape

AI
ML
DL Data
Science Statistics/
Math/EDA
Machine Learning

Supervised Learning Unsupervised Learning Regression

Game AI
Skill Acquisition
Dimensionality Learning Tasks
Clustering
Reduction Robot Navigation
Regresion Classification Real Time Decisions

Structure Discovery Recommender Systems


Advertising Popularity Prediction Feature Elicitation Targeted Marketing
Weather Forecasting Diagnostics Big Data Visualization Customer Segmentation
Customer Retention Meaningful Compression
Market forecasting
Estimating life expectancy Image Classification
Identity Fraud Detection
What is Machine Learning?
● Machine Learning is subset of
Artificial Intelligence (AI)
● Helps system to learn
automatically and get updated
from previous experience. Training data Machine learning
model

Predictions
AI vs ML vs DL

● AI: Ability of the machine to take a


decision like human AI
● ML: This is the subset of AI. The ability of ML
the machine to learn and update from DL
previous experience.
● DL: The subset of machine learning. The
deep learning (DL) algorithms are inspired
by the functions of human brain called
artificial neural network.
Overview of ML project

Data Feature Exploratory Train the


Collection Engineering Analysis model

Test the
Deployment model
Motor Control

Charging System

Role of AI in EV
System BMS

Vehicle Control
Applications
Applications
Motor ○ Primarily controls motor speed, torque and
orientation
Control ○ For Feedback most of the techniques require
speed as feedback, In sensorless techniques it
can be estimated in following ways:
Advantages
Advantages
Motor ○ Reduce the cost of transducer and provide
more compactness with less amount of
Control ○
maintenance.
Increase the system noise immunity and reduce
the transducer cables.
○ Suitability for the harsh environment including
the temperature variation.
○ Improve the overall system reliability and
robustness.
Applications
Applications


Charging ○
ML Based Coordinated Charging
Coordinated charging improves the overall
efficiency of the power grid as it avoids
Systems distribution system overloads, increases power
quality, and decreases voltage fluctuations.

EVs integration into distribution


system can be improved when
EVs charging at off-peak period
this called coordinated
charging
EV Charging - AI Applications
Functions|Challenges?
Functions

○ Cell voltage measurement


○ Battery Temperature, Current and Voltage
measurement
○ Estimation of SOC and SOH
○ Fault Analysis and Diagnostics
○ Communication with on-board systems such as
BMS charger
○ Active / Passive Cell balancing

AI Application

○ ML based model for SOC and SOH estimation


○ Data Driven Model for Battery
Battery Management System
Battery Management System
AI Applications in BMS

Battery Degradation Model

TNOC - Total No. of Cycles ; TDOD - Total Depth of


Discharge
AI Applications

Applications

○ EMS is responsible for switching between IC


engines and motor in the HEV

Energy
Management
Strategy For
HEV
Functions

Functions

○ Torque vectoring system actually replaces


ABS,ESP and other stability functionality in EV
○ AUTOPILOT features

Vehicle Control
Autopilot - End-to-End Deep Learning

Source: Avelab-Kaist
Simplified AV System Architecture
Sensors
Ultraso
Camera Radar Lidar
und

AV Software Stack
Perception Planning Controls

Actuators
Acceleratio
Braking Steering
n

21
Thanks

You might also like