IoT in Automotive Industry

The advent of integrated circuits paved way for SoC (System on Chip) fabrication architecture which in turn gave rise to Smartphones, Miniature devices and has disrupted and transform the way we live and do business. IoT building on top of this has already triggered a new digital revolution by integrating seamlessly with electronic things in the surrounding environment resulting in an increasingly connected life.

Vehicles being manufactured currently have many connected devices in them to give a cohesive experience while also guaranteeing better safety and remote control of the vehicle. Companies like Tesla and BMW have Electric cars with remote functionalities to control Ignition, Windows, lights, trunk and almost everything can be manipulated using a smartphone while the rider is busy elsewhere. While it can be argued on the security of such interconnected technologies, today’s smartphones have state-of-the-art security and multi-level encryption features which are even hard to tamper by government agencies.

• Connectivity in the car is established using persistent 3g / 4g, Wi-fi and Bluetooth. This is required by other connected devices and sensors to stream data to Cloud.
• Head mounted, side mounted and rear mounted cameras, Radar and Ultra-sonic sensors for Automatic Parking, Collision detection, Active Cruise control and speed limit monitoring, Autopilot. The data from all these sensors are then ran by against an AI model running locally which takes necessary decisions during autopilot mode.
• Vehicles can detect the presence of Key FOB (controlling handheld) within a certain distance from the vehicle’s perimeter.
• Virtually all aspects of the car functions are digitized and theoretically available for inspection in real-time or archived data on cloud including energy-use, position of wheels, brakes and emergency-brakes, climate system, seat positions, mirror positions, door handles.
• The vehicle can be ignited or woken up from sleep to ON state over the air remotely, horn and lights can be turned on from smartphone for external signalling in case of distress situations.
• The Vehicle can apply over-the-air (OTA) software updates which enhance the existing functionalities seamlessly at no extra cost just like in smartphones. Such OTA updates could be used to improve the AI Machine Learning algorithm from the server side or enable or improve a feature with existing hardware already present in the vehicle.
• Send GPS co-ordinates and seek assistance from public like Hospital or request for an ambulance or police authorities in case of an emergency in a far-off remote location.
• Monitor Driver’s riding habits, acceleration, speeding, braking efficiency in real-time and provide suggestions to improve driving guaranteeing more safety. Sensors that can track the eyeballs of the driver can detect if the driver is not fully focused, inebriated or sleepy while driving and take corrective actions by automatically parking or seeking assistance from an external party in the situation.
• On-board diagnostics and collection of data from the vehicle and generating a monthly summary of different parameters highlighting the parts / devices needing immediate replacement or attention.
• Climate-system tries to understand the weather outside and alter the auto-pilot functionalities like during a rainy weather as sudden brakes can cause the vehicle to skid therefore, auto-pilot and active cruise system would monitor the speed and apply brakes accordingly. Icy weather can cause the engine to get damaged if not turned on for a longer duration. Vehicle can sense this and start the engine autonomously or remotely with driver’s approval to keep the engine warm in such weather.

Autopilot is an advance driver assistance system comprising of lane centering, self-parking and adaptive cruise control, ability to automatically change lanes, navigate autonomously on limited access freeways, ability to summon the car to and from the garage or parking spot. It adjusts the speed and position of the vehicle by identifying the surrounding vehicles using an array of sensors in 360° direction. The latest vehicles employ Radar and ultrasonic sensors instead of camera + Laser sensors which can operate efficiently in clear weather conditions only, whereas Proximity sensors employing Radio waves can operate in all types of weather even during least visibility.

Also, companies like Tesla employ Trained Machine Learning models within the vehicle which gathers data on the Driver’s behaviour and approach of a curved road, turnings, sharp cuts, speed management, braking and so on. This data which is collected is then used to optimize and further train the M-L model by external re-inforcement from the driver during correction as can be seen below. When the auto-pilot is disengaged, the M-L model learns from the Driver and when the Auto-pilot is engaged, the Driver provides external input called Correction stimulus which is used to update the M-L Model and then finalize the optimal point of curve while travelling on freeways.

Vehicles with 360° sensory capabilities can automatically park the vehicle in the garage or parking spot by manoeuvring the vehicle in tight spots by using data from all the sensors shown in above image. It uses a combination of Proximity cameras and Radio waves-based Sensors for collision detection, environment mapping and Blind spot detection and then guides the car autonomously to park in the designated spot without any damage. These sensors are also used to lower the speed and brake if necessary, in pre-emptive collision detection system.

Temperature, humidity and Ambient sensors are employed in the vehicle to find the weather outside the vehicle during long range travelling and the vehicle’s internal environment is then adjusted based on the outside weather conditions. Ex: If the weather is very cold outside, Seat and Steering wheel heaters are automatically enabled to regulate the internal temperature and Autopilot is adjusted during rainy and Icy terrain to apply brakes and manage speed even critically when compared to clear weather conditions. The AI Model constantly monitors the data collected by these sensors and applies necessary actions to regulate the internal conditions and alter the vehicle manoeuvring based on weather and road terrain.

IoT can rapidly transform the Automotive industry and make the dream of truly autonomous vehicles into a near-future reality. AI and Machine Learning algorithms are continuously observing the Driver behaviour and approach of driving and simultaneously improving, correcting the Training Model to provide a more streamlined and tailored support to the driver when needed. This can be complemented by the Driver instilling re-inforcement or corrective actions improving the algorithm substantially. With IoT deeply embedded in the Automotive fabric, there would be truly Driverless Cars with 0 Collision and Damage rates improving the Human safety on the road.