Next-Generation MEMS IMUs — High Performance, Scalable
To find out what’s ahead for MEMS automobile navigation systems, I interviewed Yang Zhao, CEO, and Teoman Ustun, VP of Marketing and Business Development, ACEINNA, Inc. (Boston, MA).
Tech Briefs: Why do you call ACEINNA IMUs next generation?
Yang Zhao: In ideal conditions, you can navigate a car with one camera. But to achieve all time safety, and availability, autonomous vehicles increasingly rely on additional sensor technologies. However, those sensors like Vision, Radar, Lidar, GNSS, RTK can all be blinded under certain situations. IMU is the only sensor that continues to operate even when all the other sensors fail. Our challenge is to increase the amount of time for which dead reckoning with the IMU will be reliable. The current product we have is more like a high-end industrial level but designed to be incorporated into passenger cards.
Teoman Ustun: The revolutionary MEMS technology that we’re developing will achieve the performance of a fiber optic gyro at the price point of a silicon MEMS solution. That is basically a game-changing technology.
With each generation we have been improving the performance of our IMU in terms of Angular Random Walk (ARW), bias instability, and vibration/shock immunity. With this latest generation we also added a triple redundant architecture, which is extremely important for two reasons. It improves automotive level reliability and also increases the performance of the IMU. For example, to my knowledge, we have the only ASIL-B ISO 26262 (Automotive Safety Integrity Level) qualified INS (inertial navigation system) integrated with triple redundant IMU on the market. This qualification basically gives automotive customers the confidence that our sensor will generate data robustly over the life of the vehicle.
Our newest IMUs also have open architecture features that enable our customers to leverage their, know-how or to implement their secret sauce in our hardware.
Zhao: Aerospace grade IMUs have excellent stability but are large, heavy, and expensive — see Figure 1for a comparison of the different IMU grades. We are aiming to bring to the market an IMU system that will approach navigation grade. Today, you can get navigation grade IMUs, but they all use fiber optic gyros and are in the tens of thousands of dollars price range. They’re big, heavy, and expensive, but in avionics and military, that’s not a deal-killing factor.
But when you talk about using an IMU in a car, nobody would pay that kind of money. So, what we’re trying to do is to develop a new generation MEMS IMU technology that can reach navigation grade performance, which can approach fiber optic gyro performance levels. Unlike others who initially targeted consumer-grade IMUs and now are trying to shift them to higher performance applications, we started with high-performance industrial grade applications in mind, to set the requirements for our automobile sensor.
The most recent automotive products that we introduced are cost effective. They are very much designed and developed from the bottom up to have a price/performance optimization for high volume applications. We are aiming for a price of less than $500 and eventually, even less.
Tech Briefs: Could you explain some of the features that improve the accuracy, stability, and reliability of your new IMU.
Ustun: First of all, our IMU sensors have triple redundancy — there are three accelerometers and three gyros. We monitor all the critical parameters, including voltages, currents, and temperatures, as well as the performance of individual IMUs. Having this triple redundancy enables us to implement continuous self-testing. We obtain good confidence about the gyro and accelerometer performance by comparing and contrasting this triple data stream.
We also have six degrees of freedom: the three accelerometers and three gyros measure the XYZ axes. The gyros measure rotational speed, which when integrated, gives change in angular position. The accelerometer measures acceleration, which when integrated gives velocity and when integrated again, gives position. By processing gyro data and accelerometer data, and using filters and algorithms optimized for a given application, we get really accurate angular data for roll, pitch, and yaw (See Figure 2). The most critical data for an automotive application is yaw, which indicates the degree the car is drifting. Pitch indicates whether you are going up or down a hill. Hopefully, you will never see a roll in your car.
Tech Briefs: What kind of accuracy could you sustain over what time period?
Ustun: That’s very difficult to say because it depends on so many different scenarios. Currently we have to achieve lane level accuracy. A lane is about three meters wide, and a passenger car is about two meters wide. Since you want to make sure that you’re not going beyond the lane, you have about plus or minus half a meter before you’re drifting too far. At typical automobile speeds, a correction of half a meter could be achieved in less than a second. Since, under dead reckoning, you can keep the required accuracy for 60 to 90 seconds, our IMU can easily keep a car within its lane.
Tech Briefs:So, an IMU gives you relative accuracy. What about absolute accuracy?
Ustun: Absolute location data is generally derived from a GNSS signal. Historically, the positional accuracy of GNSS receivers has been a meter or more. So, GNSS by itself is not good enough for the required automobile lane-level accuracy of down to 2 cm.
Real-Time Kinematics (RTK) is a technique to improve the positional accuracy of a GNSS receiver. It uses a network of base stations that can send corrections to our vehicle-mounted RTK positioning engine, which uses them to recalculate its position more accurately.
RTK is a service that you can buy or subscribe to from many sources. What we’re seeing more and more is communication infrastructure — even telecom service providers like Verizon, Softbank, or China Mobile — have started offering RTK services. From our product perspective we are agnostic to the service provider — we intend to support any and every one of those RTK service providers to help our customers achieve the accuracy limits they’re looking for. We’re providing a solution that’s both hardware and software.
Although we do not provide the service, we enable our customers to choose the one that’s most suitable to them. However, we do qualify those services first. We make sure that they are compatible with our hardware and software. And then we put it up on our website to give customers the flexibility to choose from a list of service providers. Most commonly, RTK comes as a service over a 4G LTE network. If there is no 4G LTE reception, there are alternative ways to receive the correction signals, such as satellite-based communication.
For RTK, you put base stations across a network: there’s a rule of thumb that usually you put base stations on a 50-mile radius. These base stations are positioned so that their locations are known very accurately. Corrections are generated by measuring the error between the GNSS signal and the accurately known location of a base station.
So, we have a comprehensive solution: An inertial navigation system that includes our triple redundant high performance six degrees of freedom IMU coupled with a dual frequency GNSS receiver, in two different form factors. One is a module form factor. If the customer wants to reflow that in their own Engine Control Unit (ECU) they can do that, or otherwise they can use our packaged plug and play solution. Basically, you just connect the GPS/GNSS antenna, and the system will run on CAN bus or Ethernet.
This article was written by Ed Brown, Editor of Sensor Technology. For more information, visit here .
INSIDERElectronics & Computers
MIT Report Finds US Lead in Advanced Computing is Almost Gone - Mobility...
University of Rochester Lab Creates New 'Reddmatter' Superconductivity Material...
INSIDERElectronics & Computers
Airbus Starts Testing Autonomous Landing, Taxi Assistance on A350 DragonFly...
Boeing to Develop Two New E-7 Variants for US Air Force - Mobility Engineering...
PAC-3 Missile Successfully Intercepts Cruise Missile Target - Mobility...
Air Force Pioneers the Future of Synthetic Jet Fuel - Mobility Engineering...
How to Maximize the Benefits of Medical Device Onshoring
Leveraging Machine Learning in CAE to Reduce Prototype Simulation...
Driver-Monitoring: A New Era for Advancements in Sensor Technology
Manufacturing & Prototyping
Tailoring Additive Manufacturing to Your Needs: Strategies for...
How to Achieve Seamless Deployment of Level 3 Virtual ECUs for...