In the past few years, artificial intelligence-driven equipment and services have been the most popular and interesting items at global technology fairs. In addition to this, there are also obvious new trends in the market in 2018. First, mainstream consumer products are increasingly integrating artificial intelligence, not just those high-end and advanced products. Second, many products have artificial intelligence built into the front-end equipment instead of relying on the cloud. These trends are improving the accessibility, privacy, and efficiency of smart applications, making them more widely adopted. The world’s largest consumer electronics and technology trade show is just around the corner. Let’s take a look at the eye-catching innovations and interesting products and the latest technologies that will drive them this year.

　　 The power of 3D vision: from companion robot to deconstructed SLR camera

　　 Buddy is a home assistant robot that can navigate and move at home. It won the CES Best Innovation Award (robot and drone category). This robot is equipped with 3D vision and multiple sensors that can assist, educate, entertain and connect users, as well as monitor safety when there is no one at home.

　　 Buddy's embedded hardware supports multi-mode navigation

　　 The technology that supports Buddy’s navigation skills is real-time map creation and positioning, automatic collision and barrier prevention, and remote control by users through the smartphone screen. The exciting thing is that robots have the potential to learn new skills, because the SDK and most of the software are open source, allowing anyone interested in developing applications and skills. And Amazon's Alexa and other virtual assistants have just begun to add functions with the help of screens and cameras, and robots may be the next evolutionary direction.

　　 Another winner in the smart city category is Amaryllo AR4x. The camera robot uses embedded AI to recognize faces and other objects and react accordingly, including opening doors for familiar faces or alerting potential intruders. It has a 1080p resolution, night vision function, can receive power via Ethernet, is simple to install, and consumes limited power.

　　 Use multiple small lenses and computational photography to flatten SLR cameras

　　 The camera quality of smartphones has recently been greatly improved, most of which is due to the integration of computational photography, deep learning and dual cameras in new phones. Light took these technologies to the next level in their L16 cameras, which won the Best Innovation Award in the CES Digital Imaging category. Unlike SLR cameras that use a large lens and sensor, or smartphones with one or two small cameras, the company uses an array of 16 small camera modules, each equipped with a lens, sensor and mirror.

　　 Every time the camera takes a photo, at least 10 different modules take the same image from a slightly different angle and different focal length. Then the camera uses intelligent processing to stitch all the images together to generate a high-resolution image. By calculating the subtle differences in each picture and the angle of the lens, a depth map is constructed for the shooting scene. This 3D information is very helpful for many post-processing functions, such as changing the depth of focus. By making a deconstructed SLR camera into the shape of a portable smartphone, the best of both worlds is achieved as much as possible.

　　 Semi-autonomous cars enter the mass market with efficient sensors and processors

　　 Tesla’s price tag allows only a limited number of people to enjoy safe and convenient autonomous driving functions. However, Nissan chose to target the mass market, offering their ProPilot technology on most models rather than the most expensive models. Leaf electric vehicles and Rogue compact SUVs, which won the Best Innovation Award in the CES Vehicle Intelligence and Autonomous Driving Technology category, will be launched in 2018 and will be equipped with ProPilot technology.

　　 ProPilot technology uses a combination of multiple sensors, including vision and radar, to perform intelligent cruise control and lane keeping assistance when driving on a single-lane highway. When clear lane markings are visible, the control module can drive the vehicle, but the driver's hands must be kept on the steering wheel to operate the system.

　　Nissan integrates ProPilot assistant on best-selling models to push semi-autonomous driving to the mainstream

　　Smart Voice: Personalized immersive audio and in-ear simultaneous interpretation

　　 The nuraphone, which won the best innovation award in the earphone category, is an interesting product. They look like headphones, but there are silicone earplugs inside that can be inserted into the ear canal. This design is very effective for isolating passive noise. It allows the melody to enter the ear directly, while the bass line uses bone conduction technology to pass through the skin. These features are designed to produce a highly immersive listening experience, just like a live performance.

　　 The interesting combination of headphones and earphones provides immersive audio

　　 Another feature of Nuraphone is that they can learn and adapt to each user's unique hearing. When used for the first time, the headset plays a specific frequency, prompting the user's ear to generate a subsonic sound called otoacoustic emission (OAE) from the cochlea. Nuraphone has a very sensitive microphone to monitor OAE, and a built-in self-learning engine to decode the sensed sound information and create a personalized auditory file.

　　 This is a very innovative headset, but more exciting technologies can be integrated into this type of device. Many new headsets and ear-worn devices have active noise cancellation, voice activation, and even voice augmented reality functions. For example, the Mars headset, another product that won an award in the headset category, includes these features. These are true wireless stereo headsets that support real-time translation in ten languages and do not require any external equipment.

　　 One of the biggest challenges is how to combine these features with a battery that can last all day. This requires ultra-low-power microphones and processors, and appropriate standards to transmit audio streams on Bluetooth Low Energy (BLE).

　　 Looking forward to meeting at CES 2018

　　 CES will exhibit many exciting technologies and products. We will definitely see impressive AR/VR technologies, drones and more. But the award-winning products mentioned here indicate that artificial intelligence, machine learning and complex algorithms are entering daily life and are practical for most people. Many of these devices are only equipped with small batteries or other limited power sources, so energy efficiency becomes a key factor that determines how these devices operate better and how often they need to be recharged. With the advancement of these technologies, we will see that more cameras, sensors, and microphones are integrated into such products, and power consumption will become more important. These technologies are accompanied by high-efficiency and low-power processing, and will have unlimited development opportunities.