A Single-Package Proximity-based Gesture Recognition Sensor for Mobile Applications
- 주제(키워드) Gesture Recognition Sensor
- 발행기관 고려대학교 대학원
- 지도교수 김석기
- 발행년도 2014
- 학위수여년월 2014. 2
- 학위구분 박사
- 학과 일반대학원 마이크로/나노시스템협동과정
- 원문페이지 120 p
- 실제URI http://www.dcollection.net/handler/korea/000000050099
- 본문언어 영어
- 제출원본 000045793187
초록/요약
Recently developed smart phones and tablets toward various applications including healthcare, monitoring user status, and motion recognition. This trend requires various human-machine interfaces (HMIs) that operate under various circumstances and seamlessly combine multiple sensors such as a touch screen, gyroscope, accelerometer, digital compass, camera, and proximity. Gesture recognition sensors (GRSs) are one of the emerging research topics as intuitive and natural interfaces for portable handheld applications, which requires these sensors to be small and inexpensive while consuming low power. Current HMIs can be categorized into four different types: touch-based, motion-based, vision-based, and proximity-based systems. Touch-based systems are intuitive but users have to physically contact the screen for activation. Motion-based systems support single handed interaction by holding a device and moving it. Accelerometers or gyroscopes are used to measure inertial movement at the expense of heavy computation burden. The vision-based systems utilize an embedded camera that detects user’s contactless motion. However, massive computations are required, resulting in increasing power consumption. For overcoming these problems, proximity-based system was introduced. In active infrared (IR) proximity-based GRSs using multiple light emitting diodes (LEDs), IR LED alternately reaches onto a PD after reflected from an object. Two alternated electrical signals from the PD vary by the locations of the object, which allows the gesture recognition algorithm to detect various motions. In this case, the distance between LEDs determines the recognition rate. Since LEDs are the most power-hungry components in the system, a GRS with a single LED was introduced to reduce the overall power consumption. This configuration enables low-power sensor design, but the shorter the distance between PDs still causes the lower recognition rate. As a result, a large form factor cannot be avoided, which makes it hard to be applied to portable devices. To put PDs close to each other in a single chip without degrading the recognition rate, the gesture recognition sensor using an optical block was proposed. However, it requires additional structures and processes to implement the conventional optical block into a package, resulting in increasing package price. Furthermore, since that conventional proximity-based gesture recognition sensor cannot be integrated with a driver for IR LED, it is impossible to implement a single-package including an IR LED. In this thesis, in order to overcome limitations of the conventional GRSs, novel optical block and pseudo-differential based light to digital converter are proposed. The proposed optical block can confine the field-of-view of each PD, but requires no additional process for implementing. Also, since proposed pseudo-differential light to digital converter has the immunity against supply and ground bounce, the proposed GRS can be merged with an IR LED into single-package. And, by proposed noise reduction mechanisms, the proposed gesture recognition sensor can reduce noise effectively. In order to apply mobile applications, the proposed gesture recognition sensors was designed to detect more than 12 cm while consuming power less than 4 mA. And then the proposed GRS was fabricated with 0.18 μm CIS technology and tested under various conditions.
more목차
Table of Contents
1. Introduction 1
2. Conventional Human-Machine Interfaces 4
2.1 Touch-based gesture recognition sensor 6
2.1.1 Projected capacitive touch-based System 7
2.1.2 Resistive touch-based system 9
2.2 Motion-based gesture recognition sensor 11
2.2.1 Motion-based GRS with Accelerometer 12
2.2.2 Motion-based GRS with Gyroscope 13
2.3 Vision based gesture recognition sensor 15
2.4 Proximity-based gesture recognition sensor 17
3. System overview and analysis of proposed optical structure 20
3.1 System overview of proposed gesture recognition sensor 22
3.2 Analysis of proposed optical block 24
3.3.1 Conventional optical block 24
3.3.2 Proposed configuration 26
4. Light to digital converter 28
4.1 Conventional light to digital converter 30
4.1.1 Integrator based light to digital converter 30
4.1.2 Single-ended based light to digital converter 32
4.2 Proposed light to digital converter 36
4.3 Design of the proposed light to digital converter 43
4.3.1 Fundamental noise mechanism 43
4.3.1.1 Thermal Noise 44
4.3.1.2 Low frequency noise 46
4.3.2 Ambient light noise 47
4.3.3 Photodiode 49
4.3.4 Trans-impedance amplifier 51
4.3.4.1 Resistive TIA 52
4.3.4.2 Common-gate TIA 53
4.3.4.3 Resistive Feedback TIA 56
4.3.4.4 Summary of TIA’s properties 58
4.3.5 Pre-amplifier 60
4.3.6 Programmable Gain Amplifier (PGA) 66
4.3.7 Correlated double sampling circuit 68
4.3.8 Analog to digital converter 72
5. Simulation Results 75
5.1 Simulation results for basic operations 75
5.2 Simulation results for a swipe from right to left 80
5.3 Simulation results including supply/ground noise 83
5.4 Frequency response of the pre-amp 85
6. Fabrication and Test 86
6.1 Chip fabrication 86
6.2 Test results 89
6.2.1 Measured outputs of the gain stage 89
6.2.2 Maximum detection distance 92
6.2.2.1 White card with 2 cm width 92
6.2.2.2 Gray card with 2 cm width 94
6.2.2.3 White card with 8 cm width 95
6.2.3 Performance of the proposed optical block 97
6.3 Comparison results 99
7. Conclusion 100
References 102
