You may not have heard of Piezomems, but novel applications of this game-changing technology are ready to reshape the future of AI.
Researchers estimate in 2023 that using Generative Artificial Intelligence (Genai) creates images of as much energy as smartphones charge. Now, imagine using a smartphone to generate AI images.
As AI moves to edge devices like smartphones and AR glasses, the most eye-catching and widest range of applications may spin around real-time, personalized, and context-aware experiences. These devices are always with us, so AI can leverage its sensors – cameras, microphones, GPS, accelerometers – to provide seamless, low latency solutions and experiences.
Think of AI as an omnipresent assistant, providing real-time contextual help. On a smartphone, this could mean instant language translation during conversation or travel, such as pointing the camera to a sign and covering it in your native language. For Ar glasses, it’s even more shocking – imagine walking through the city and projecting historical facts, restaurant reviews or navigation tips in your field of view, all of which fit your interests without unplugging the device. Since Edge AI processes data locally, this is a win for trust and data privacy.
While Edge AI’s promise is obvious, there are still hardware challenges.
- Challenge 1 – Thermal: Just like AI data centers, Edge AI devices are thermally limited, not MIPS/Compute limited. Smartphones are already under the limit of popularity; Edge AI capabilities simply complicate the problem. Likewise, with more computation, optical and micro displays.
- Challenge 2 – Size/Weight/Exterior: This is a huge challenge for AR glasses as manufacturers seek to balance the comfort of all-day wear (weight) with the holy grail of necessary electronics and battery life (performance), all in fashion-first, stylish form factor.
- Challenge 3 – Conversation AI Audio Quality: Sound will be a major interface, especially in the case of AR glasses, which can operate quickly and hands-free. However, traditional coil speakers are thick and bulky (see Challenge 2) and do not reach their maximum potential in tight, constrained spaces. Furthermore, their relatively poor partial consistency negatively affects value-added DSPs such as privacy and conversation focus patterns.
Still, the crazy dash (smartphones, AR glasses and other mobile devices) that execute Genai on edge devices is still in progress. Deloitte researchers estimate that by the end of this year, Genai-enabled smartphones could exceed 30% of the shipment. Needless to mention the new Genai-powered smart glasses.
In edge devices, some of the solutions come from reimagining the electronics that enter them.
Piezoelectric components enable Genai at the edge
Piezomems technology addresses the challenges of Genai in edge devices. Not only because of its function, but also because of how it is made.
Piezomems is an application of micro-motor electromechanical system technology, which uses piezoelectric materials to convert electrical energy into motion, essentially a voltage-controlled source of airflow. Piezoelectric components are manufactured in a reliable, stable and highly uniform semiconductor process, allowing efficient, cost-effective microelectronics to be mass-produced in small 100 mm thin chip packages.
Active cooling of EDGE AI thermal management
There are now ultrasonic piezoelectric silicon actuators that can draw out completely silent vibration-free pulses, with the aim of cooling thin, space-constrained electronic systems – basically fans on the chip.
The impact on cooling edge AI edge devices is far-reaching. The first Piezomems micro-cooling fan is 96% smaller than a traditional fan and is the only active thermal management device, the only small and thin enough to fit inside smartphones and AR glasses, keeping surface and component temperatures 15-30% cool while allowing computing to run at peak performance.
Piezomems speakers serve as dialogue AI interface
Size and Weight: Piezoelectric speakers can produce equivalent or better loudness (sound pressure level [SPL]) As a conventional coil speaker, but a small part of size, thickness and weight. Applying this new speaker to AI-enabled glasses can bring product designers closer to the Holy Grail balance cited earlier in this article.
These are two examples. First, at 1mm thin, Piezomems speakers are about 70% thinner than traditional coil speakers, which makes glasses’ arms thinner and more stylish.
Second, it is suggested that the ideal weight of the glass is 30 grams for comfort throughout the day. In AI glasses, traditional coil speakers usually weigh about 2 grams. There is one in each glasses arm (4 grams in total) and the speakers make up about 15% of the total weight of the system. Each of the piezoelectric speakers is about 150 mg, which eliminates 90% of the weight caused by the speaker, bringing Ar glasses closer to the 30 g target.
Audio clarity. In addition to its small size and weight, Piezomems speakers also enhance the audio experience. Not only do they produce enough volume in the open air, they also perform well in replicating high frequencies that are often associated with improving speech clarity and clarity. This is because Piezomems are mechanically responsive faster than traditional speaker designs with almost no phase shift, which means the audio is clear, detailed and accurate.
Improved DSP functionality. Finally, DSP-intensive features such as privacy mode and directional sound will perform better with Piezomems speakers. Due to the uniformity and consistency of semiconductor processes, these speakers have nearly ideal part loudness and phase matching, providing a more predictable window for the DSP algorithm. With less variability in speakers, DSP algorithms can improve performance with less processing overhead.
The potential of unlocking edge AI devices
Finally, it takes a lot of innovation to bring the power of AI to consumer edge devices. Manufacturers must not only overcome physics, but also ensure the best user experience – all without compromising the form factor.
New innovations in Piezomems technology open up the possibility of eliminating thermal constraints, enhancing speech clarity in style (thin) and comfort (thin) and comfort (weight) in Edge AI smartphones and AR glasses.
The Genai ecosystem is vast and interconnected. Piezoelectrics can be the key pusher.
