Can an EV Understand Its Own Health?

Electric vehicles are becoming increasingly common around us.

Whether it is electric two-wheelers for daily commuting or electric cars for long-distance travel, the transition toward electric mobility is accelerating every year.

But while EV technology has evolved rapidly in terms of batteries, charging infrastructure, and driving experience, I often feel that one important area receives less attention than it deserves:

understanding the actual condition of the vehicle while it is operating.

Most EVs today provide information such as:

• Battery percentage

• Remaining range

• Charging status

• Basic warnings

These indicators are useful.

But they made me think—

is that enough?

Because if we think about it carefully:

A battery showing 65% charge does not tell us:

• whether the battery is aging abnormally,

• whether the charging cycles are affecting efficiency,

• whether vibration is slowly damaging electrical interfaces,

• whether connector degradation is increasing losses,

• or whether the vehicle can still support communication during an emergency.

These questions became the starting point of this project.

Looking Beyond Battery Percentage

As users, we generally evaluate an EV through two numbers:

Battery Remaining
Estimated Range

But internally, an EV battery system is continuously experiencing:

• charging cycles,

• discharge patterns,

• temperature variation,

• current spikes,

• environmental vibration,

• connector wear,

• changing efficiency.

Many of these changes happen gradually.

By the time symptoms become visible to the rider—

performance may already be affected.

That raised a question:

Instead of simply reporting battery percentage—

can a vehicle become more aware of its own condition?

Can it provide meaningful insight before problems become visible?

That idea became the foundation of EV Guardian.

Introducing EV Guardian

Intelligent EV Battery Insight, Safety & Emergency Response System

EV Guardian is an attempt to build a distributed intelligence platform around three major pillars:

Note: AI Generated image for reference.

Understand energy.

Understand safety.

Maintain communication during emergencies.

Instead of treating the vehicle as a single black box, the architecture separates responsibilities into dedicated functional nodes.

This decision was intentional.

Battery intelligence and emergency response solve different problems and should remain independently operable.

The current architecture consists of two primary nodes.

Node 1 — Battery Management & Profiling Node

This node forms the energy intelligence layer of the system.

Its responsibility is not limited to measurement.

Its objective is to observe battery behavior and extract meaningful insight.

The subsystem continuously monitors:

• battery voltage,

• battery current,

• charging state,

• discharge behavior,

• thermal conditions.

Alongside monitoring, it also manages protection mechanisms:

• overvoltage protection,

• undervoltage protection,

• overcurrent cutoff.

At this stage, these are standard battery management capabilities.

But the interesting part begins after measurements are collected.

From Measurement to Battery Insight

Most systems answer:

What is the battery level?

This node attempts to answer:

What is happening to the battery?

Using edge-side analytics running locally, the system will investigate:

State of Charge (SoC)

Estimating usable remaining energy.

Not simply:
“60% remaining”

But:

“How much operation time remains under current conditions?”

State of Health (SoH)

Estimating long-term battery condition.

This provides visibility into:

• degradation,

• available capacity,

• expected performance.

Battery Anomaly Detection

Battery failures often develop gradually.

Early indicators may appear as:

• unexpected current patterns,

• abnormal charging profiles,

• efficiency reduction,

• thermal drift.

The objective is to identify patterns before they become user-visible failures.

Predictive Maintenance

Rather than waiting for issues to appear—

the system attempts to estimate when intervention may become necessary.

The long-term goal is moving from reactive maintenance to informed operation.

Node 2 — Safety & Emergency SOS Node

While battery insight improves reliability—

safety remains equally important.

This node acts as the central coordinator and safety controller.

It continuously monitors vehicle dynamics using inertial sensing.

Measured conditions include:

• acceleration,

• abrupt motion,

• vibration,

• impact events.

The intention is not to classify accidents immediately.

Instead, the goal is to establish a framework capable of recognizing abnormal operating conditions.

Future software iterations may investigate lightweight edge models for better motion interpretation.

Safety Should Continue Even After Failure

One scenario I kept returning to while thinking about this design was:

What happens if the vehicle becomes unavailable during an emergency?

The rider may still require:

• communication,

• location sharing,

• emergency assistance.

This inspired another subsystem.

A small redundant battery path is planned for critical operation only.

This is not intended for vehicle propulsion.

Its only objective is preserving essential functions.

Examples include:

• emergency SOS triggering,

• phone charging,

• location transmission,

• basic system diagnostics.

The philosophy here is simple:

If transportation stops—

communication should not.

Why Use Distributed Communication?

The two nodes communicate through isolated RS485.

This choice was intentional.

Transportation systems operate in electrically noisy environments.

Separating functions while maintaining reliable communication offers:

• fault isolation,

• modular expansion,

• improved robustness,

• cleaner architecture.

Battery intelligence remains independent.

Safety intelligence remains independent.

But both continue exchanging critical information.

Making Vehicle Health Visible

All collected information is transmitted wirelessly to a visualization interface.

Initially this will run on a laptop.

The dashboard will eventually display:

Battery Layer

• State of Charge

• State of Health

• Estimated range

• Efficiency indicators

Safety Layer

• Vibration profile

• Warning states

• Emergency events

System Layer

• Node status

• Communication health

• Event history

The current interface is intentionally developed to remain portable to future embedded displays.

Closing Thoughts

This project is not trying to redesign electric vehicles.

It is exploring a smaller question.

As EVs become more intelligent—

should they simply display status,

or should they actively help users understand risk?

Battery insight.

Predictive diagnostics.

Emergency preparedness.

Resilient communication.

These may become increasingly important components of future mobility.

This is the beginning of EV Guardian.

Next, I will move from concept into implementation and begin building the distributed architecture behind the system.