Automotive

10 things to look out for when choosing EV software

20th April 2021
Lanna Deamer
0

Tech innovation is rocking the automotive world. As Electronic Vehicles (EVs) make headway, autonomy, connectivity and data-sharing are the next big steps for EV makers looking to create responsive, modern vehicles. Mobility leaders must choose their EV software providers carefully, and should consider the following critical trends.

By Jörg Zimmer, VP EMEA Sales, BlackBerry Technology Solutions

Great software is a must-have: Electric cars rely on many features where software predominates - Advanced Driver Assistance Systems (ADAS), connectivity, infotainment, battery monitoring, and attractive User Interfaces (UIs) and more.

Auto-makers must own software decisions: Traditionally, outsourcing has been the way with vehicles. But with software so critical to EVs, the automaker must now design and own the car’s software so it can be built, fixed, extended, and continually maintained. 

Software-first is the best approach. Traditional automakers don’t yet have enough software skills and vice versa: tech companies getting in on the act need automotive knowledge. EV makers should therefore choose a partner with deep automotive software expertise and successful track records to help them solve their challenges. 

Here are 10 things to look out for when choosing an EV software provider.

1. Safety first

The most critical requirements of any vehicle system is safety. 

A safe software provider is a business which makes the right investments in safety certifications and meeting industry standards for safe products. EV makers should therefore choose a software partner who helps them streamline their safety certification processes and meet aggressive start-of-production dates, through a combination of pre-certified software and unique safety expertise. 

2. Deep competence

Safety-certified software is nothing without knowledgeable engineering staff behind it. This is as true for new EV makers who sometimes lack rigorous software discipline as it is for established automakers building new lidar systems and experimenting with infotainment. 

Most automakers’ in-house EV engineering staff won’t be clued up on certification. This is why automakers should partner with software providers and trustworthy guides to produce safety-certified software through engineer training, safety assessments, and development methodologies.

3. Whole-car software

Every connected car is supported by dozens of separate subsystems that control its every aspect: door locks, windows, braking systems, lighting, engine, suspension, airbags, ADAS, instrument clusters, and much more. A single software architecture is vital to holistically manage the software across every module - both safety-critical under-the-hood modules, and the applications on our dashboards. 

EV software needs to match new EV architectures, enabling easy maintenance, updates and fixes. While secure software can easily support safety-critical applications like engine control, it also needs to adapt to managing the modern protocols and high-powered graphics needed for smartphone connectivity and telematics.

4. Hypervisors and module consolidation

Virtualisation is another architectural change seen across the industry that is reducing costs and simplifying electronics. 

Previously all subsystems were in separate Electronic Control Units (ECUs), each black box requiring its own hardware, microprocessor, memory and storage. Now, running many ECUs on one system is bringing added value, removing the cost and complexity of dozens of hardware boxes in the car.

Hypervisors enable this, by keeping each subsystem isolated and running disparate functions such as instrument clusters, infotainment and displays. It also enables remote updates for EV teams, and prevents catastrophic cyberattacks through isolating each subsystem.  

5. Built for security

When cars weren’t connected, there was little worry about remote attackers stealing data or disabling cars. That’s no longer the case; new, always-connected and software-heavy EV architecture must be designed for defence.

Automotive security is facing an overhaul. The UN’s WP.29 regulation establishes new cyber security requirements for vehicles sold in the EU, the UK, Japan, and South Korea. Its demands over cyber security process, risk mitigation strategy, cyber-safe vehicle design, and certification are set to impact automotive cyber security programs worldwide. Being able to lean on trustworthy sources of knowledge will be critical for compliance and safety.

6. Cloud connected

Everyone regularly relies on Over-The-Air (OTA) software updates for their phones and connected devices, enabling bug fixes and security patches for optimal customer use. The same is necessary in EVs, increasing perceived value and generating coveted brand loyalty. 

A good software provider helps automakers deliver an automotive-specific OTA experience - and this means non-intrusive. Phone updates can produce many minutes of unavailability as the system rewrites critical software components. Nobody wants to wait for their car to update while they’re trying to get their child to school on time. Updates must be fast, convenient and bring clear value.

7. Sensor fusion

Sensors continually collect data from autonomous vehicles and their surroundings: camera, lidar, GPS location and more. Sensor fusion then merges these continuous data sources into one data stream which drives the autonomous car’s recognition, decision, and control mechanisms. Once fed to the cloud, that vital data also feeds the machine-learning algorithms that allow vehicles to continually improve.

To achieve this, EV software should be scalable and cloud-connected, allowing automakers to provide a consistent and secure way to read and analyse vehicle sensor data before creating responsive in-vehicle services that enhance driver and passenger experiences.

8. Reliable self-driving

An autonomous system requires software that is reliable and safe. Yet a car’s self-driving system uses different software technologies than normal automotive embedded software. 

For instance, machine learning algorithms often use the Python programming language, while machine vision systems usually need graphics processing unit (GPU) programming. However, these systems still require real-time performance and high reliability. This combination of attributes and technologies is a crucial factor when choosing a vendor. 

9. Power equals distance 

Electronics designed for Internal Combustion Engines (ICE) could always rely on a running engine to provide a steady power supply. This meant that the car’s electrical components weren’t maximally energy efficient - they didn’t need to be. 

A battery fuel source makes energy efficiency a significantly higher priority, as power must be sufficiently balanced between electronic systems and propulsion. With limited range a key customer concern, poorly designed software risks using excessive power. EV manufacturers therefore need to work with their software providers to design and develop whole-vehicle power management systems that help avoid poor range performance.

10. Time-to-market

Whether old or new, EV makers must meet regulatory requirements, and appeal to customers who may be purchasing an EV for the first time. They can’t afford the time to piece together the millions of details involved in building electric car software.

The cost of homegrown solutions and extensive integration lessons is high: at best it comes with added cost and wasted time, and at worst can lead to public mistakes and brand damage. EV makers can make much better use of their time and finances by using industry-trusted products and road-tested suppliers to get their vehicles to market faster.

Featured products

Product Spotlight

Upcoming Events

View all events
Newsletter
Latest global electronics news
© Copyright 2024 Electronic Specifier