There is a lot of buzz around electric vehicles with a huge debate going on in society on the viability of electric vs. fueled vehicles in the marketplace. At the same time, there is another big technology change that is often conflated with EV, and that is the move from human driven to autonomous vehicle systems.
This move towards self-driven vehicles will occur independently from the debate on what will be powering those vehicles (spoiler alert - EVs will win). Eventually every vehicle on the road or on the campus, from self-driving forklifts on a factory floor to cross-country buses on the highways will be self-driving to some extent, regardless of the motive force under the hood.
No precision without feedback
So how do you migrate from a manual system to an autonomous one? The first step is to automate the subsystems that can be enhanced and work your way towards a total system. In vehicles, examples of that subsystem automation include transmissions, cruise control, and traction-management systems.
These systems are driven by advanced sensor suites that communicate with the system controller to provide the critical feedback needed to control the systems involved. Transmissions have hall sensors in them, cruise control systems now use radar and other distance-detection technology, and traction control needs precise wheel-speed information.
However, to permit full autonomous operation in the real world, a vehicle needs what is called proprioception. It’s the sense living creatures have that helps them keep track of their physical presence, actions, and navigation in the real world. Humans and other animals have a variety of sensors and an organic brain to integrate the data (Figure 1), and the challenge to autonomous vehicles is to mimic all that functionality in an organic system, without many of the organic senses we take for granted.
Organic systems get direction cues from smells, the pressure of wind on the skin, and a mass of other subliminal information from a combination of senses leveraged with knowledge and memory. A robotic system must use other sensors to make up for those it cannot copy. Luckily, we can use electronic systems to detect magnetic fields, determine vectors, and work with external systems like GPS for additional support.
What does autonomous really mean?
There is an industry standard for the various levels of vehicle automation. Our mathematically-minded friends will be pleased to know the level count starts at zero, as it represents the manual baseline. Basic cruise control falls under this, as it only maintains a constant speed without braking or acceleration.
Level one and two are pretty basic (self-steer) and advanced (the car steers) versions of lane-keeping cruise control, adding the ability to detect the road and the cars around it. Levels three through five add more and more autonomy until at the top of the list you have a self-driving car you could trust to take you home at night.
Let’s look at an autonomous forklift in a facility. In order to give it autonomous operational capability, not only does it need to know where it’s wheels, lifting fork, and load are positioned at all times, it also needs to know where it is in the facility, so it can travel to its destination, find a path, and know it has gotten there.
There are some workarounds to true presence sense in a facility. One can put magnets in the floor or walls or use optical guidance markers, some way to tell the robot forklift where it is and where it needs to go. These work in a very simple and closed process but can have difficulty traveling in factories with multiple buildings or different levels.
This problem is exacerbated in autonomous vehicles in the real world, especially in challenging applications like military logistics delivery. Such vehicles can’t follow pre-positioned magnetic markers or trust local landmarks and signage for guidance. That’s why an inertial measurement unit (IMU), a subsystem that monitors the dynamically changing movements of the vehicle, is a critical enabler in an autonomous vehicle.
The IMU is a sensor module that provides data on the vehicle’s angle, direction, and speed so the system controller can manage vehicle position and trajectory, as well as detect safety issues like changes in vehicle angle.
Depending on the core technologies involved, an IMU can measure vectors, the vehicle’s angular rate, and depending on the level of functionality desired, the local magnetic fields. The IMU is a system-in-package that can track inertial movement by detecting linear acceleration and rotational rate integrating data from a combination of (or a single multi-axis device) micro electro-mechanical system (MEMS)-based accelerometers and gyroscopes.
The number of axes an IMU can handle is expressed in Degrees of Freedom (DOF), where the number represents the amount of axes covered and in which manner. A sensor with 3 DOF could be a 3-axis accelerometer or a 3-axis gyroscope, but not both. If you combine both, you get 6 DOF. Another way to get the number is by combining an accelerometer and a magnetometer for a tilt-compensated compass.
For use in the most demanding outdoor situations, like automated farming and mil/aero applications, a 9DOF IMU takes both the 3-axis accelerometer and gyroscope and adds a magnetometer to the mix to provide compass readings. For example, the ACEINNA IMU381ZA-209 carries in its small footprint a fully-calibrated 9 DOF IMU for demanding embedded applications in a robust low-profile package. The IMU has a standard SPI bus for easy system communications.
The IMU’s feedback on a vehicle’s position provides needed data to enable a navigation system to bridge situations where GPS signals or other external guidance indicators are weak or missing. By supplementing the other sensor information with an IMU, an autonomous vehicle can self-determine all aspects of its mobility and direction.
Current multi-axis accelerometers used for vehicle stability and rollover detection aren’t able to perform vehicle-presence tasks accurately. Whether from a lack of sensitivity and accuracy in the simple devices used, to incomplete sensor integration, these legacy devices aren’t quite adequate to the task.
One advantage with an IMU is that it can take charge of all vehicle stability and motion monitoring, integrating sensors and simplifying the system’s design. The IMU provides the critical information that allows the vehicle to integrate the other sensor information in such a way that it can determine all aspects of its mobility and direction.
IMUs, were once such expensive and complex devices that they were only used to serve Mil/Aero applications in both manned and unmanned aircraft and spacecraft, including satellites and planetary landers. With the latest generation of IMUs not only cost-effective, but simple to integrate, robust, and reliable, using them in your next design is a real force multiplier.
The future of the autonomous vehicle is still being written, but it is undeniably growing and expanding in many ways. One of the constants will be the need for precise position information, and the IMU, working with other sensors such as RADAR, LiDAR, cameras, GPS, etc,) is the best way to achieve it.
ACEINNA Inc., headquartered in Andover, Massachusetts, provides leading edge MEMS-based sensing solutions that help our customers improve the reliability, cost, features, and performance of their end products and equipment. The company has manufacturing facilities in Wuxi, China, and R&D facilities in San Jose CA, Andover MA, and Chicago IL. FOR MORE INFORMATION on ACEINNA Inc., One Tech Drive, Suite 325, Andover, MA 0180, Tel: 978-965-3200, Fax: 978-965-3201, Email: email@example.com, Web: https://www.aceinna.com.