Demo-1:
Demo Title: Robust Adaptive Game-Theoretic Decision-Making for Autonomous Driving
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 112+114
Demo Exhibitor:
Jinjun Shan, York university
Mingfeng Yuan, York university
Demo Description:
The demonstration is about a complex traffic scenario where all road users enter an intersection at the same time without any traffic signs or signals. The main idea behind this demo is to learn a robust and adaptive driving policy using a DRL method in a simulator and directly apply the learned model to real-world scenarios without any further training. To reduce the sim-to-real gap and to support end-to-end DRL training, we developed a high-fidelity simulator based on ROS-gazebo that can simulate sensor data and vehicle dynamics. Since we model the decision-making at the intersection as a Partially Observable Markov Decision Process, all decisions of vehicles are made based on the onboard sensors without any communication among vehicles. To model the strategic decision-making process of human drivers, a game-theoretical concept named level-k reasoning is used. Different driving behaviors (including conservative, aggressive, and adaptive policies) are trained iteratively. The lowest level of reasoning in this concept is called level-0 reasoning, which is the anchoring policy from which all the higher levels are derived using DRL. A level-0 policy is a non-strategic driver since their decisions are not based on other vehicles’ possible actions but consist of predetermined moves. To obtain the level-k policy, a traffic scenario is created where all drivers are level-(k-1) drivers except the ego car that is to be trained to best respond to the level-(k-1) policy. Once the training is completed, the ego becomes a level- k driver. Therefore, diverse interaction behaviors are successfully modeled through the proposed self-play training scheme, showing great significance in improving the safety of self-driving cars. More details can be found in the recorded video.
Video Link: https://www.youtube.com/watch?v=mPtoojXh2-s
Demo-2:
Demo Title: Holographic Radio: A New Paradigm towards Ultra-massive MIMO enabled by Reconfigurable Holographic Surface
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 122+124
Demo Exhibitor: Boya Di, Peking University
Demo Description:
In this demonstration program, we implement our own designed reconfigurable holographic surface (RHS) and build up a RHS-enabled holographic radio system where the RHS serves as transmitter antenna array. The feasibility of utilizing reconfigurable holographic surfaces (RHSs) to enable holographic radio at both the theoretical level and system level is demonstrated by fully exploiting the potential of amplitude-controlled holographic beamforming.
The developed prototype of RHS-enabled holographic radio shown in Fig.1 consists of the transmitter (Tx) side and the receiver (Rx) side. The Tx side includes a Tx host computer controlling the Tx via a software program; a Tx universal software radio peripheral (USRP) LW-N210 conducting baseband signal processing and RF modulation; and our implemented two-dimensional RHS realizing holographic beamforming. The Rx side includes Rx antennas receiving signals; a Rx USRP LW-N210 conducting RF demodulation and signal processing to recover the original signals; and the Rx host
computers displaying the recovered signals and the corresponding parameters. More precisely, we contribute to the research on RHS-enabled holographic radio from the following three aspects:
First, we present our hardware design of RHS elements with controllable radiation amplitudes. Full-wave analysis is then conducted to analyze the radiation characteristics of the RHS elements, based on which a holographic beamforming optimization algorithm is developed for beampattern gain maximization. It is proved that an RHS element loaded with PIN diodes capable of binary discrete amplitude control is enough for beampattern gain maximization.
Second, we have implemented a prototype of an RHS consisting of a two-dimensional array of multiple electrically controllable RHS elements. As shown in Fig. 3, there are 4 feeds attached to the left end of each row. The electromagnetic wave generated from each feed propagates to each RHS element and radiates energy to free space. The radiation amplitude of each RHS element can be controlled based on a field-programmable gate array (FPGA) (ALTERA AX301). Specifically, the bias voltage applied to each PIN diode (MACOM MADP-000907-14020) can be changed by controlling the FPGA, such that the ON/OFF state of the PIN diode together with the radiation amplitude of each RHS element can be controlled.
Third, to substantiate the feasibility of RHS-enabled holographic radio, we have built an RHS-aided communication platform. Experimental results have verified the effectiveness of the proposed holographic beamforming optimization algorithm. The capability of the RHS-aided communication platform to support real-time transmission of high-definition video has also been validated. Moreover, the overall power consumption of the RHS-enabled holographic radio is only 0.686 W, which is much lower than that of a phase shifter.
Video Link: https://youtu.be/d-SmWRTlP-M
Demo-3 (Canceled)
Demo Title: Performance Demonstration of RIS Configuration Algorithms using srsRAN
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 118
Demo Exhibitor:
Sefa Kayraklık, Communications and Signal Processing Research (HISAR) Lab., TUBITAK BILGEM, Türkiye
Hüseyin Arslan, Department of Electrical and Electronics Engineering, Istanbul Medipol University, Türkiye
Demo Description:
In this demonstration, the performance of the RIS-assisted wireless communication system will be demonstrated in an indoor environment, where Greenerwave sub-6GHz RIS prototype and an open source 4G / 5G emulator (srsRAN) running on a Linux based computer integrated with Software Defined Radio (SDR) will be used as base station (BS) and user equipment (UE). The real-time video will be transmitted from the BS to the UE over the RIS-assisted wireless link. Then, the optimization of the RIS configuration will be demonstrated by employing iterative and codebook based algorithms. The results show that RIS provides over 15 dB increase in RSRP when the horn antennas are utilized resulting in higher modulation and data rate.
Video Link: https://drive.google.com/file/d/1oMxVvzriZP9vDiF0jGtfdWVK9o0S1fYO/view?usp=drive_link
Demo-4:
Demo Title: Real-Time Wireless Indoor Drone Communication
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth# 104
Demo Exhibitor:
Ahmad Ai, University of Windsor
Demo Description:
The “Real-Time Wireless Indoor Drone Communication” demo features an innovative system establishing wireless communication between an indoor drone and a Raspberry Pi. Equipped with a camera, ultrasonic sensor, and temperature sensor, the drone collects real-time data during flight, transmitting it wirelessly via a custom protocol to the Raspberry Pi.
The Raspberry Pi acts as an intermediary, processing the drone’s incoming data and publishing it to a ground station (laptop) over a local network. The station presents the sensor data in an organized and visually appealing manner on a user interface.
This seamless data transmission enables real-time data retrieval and analysis, showcasing the potential of wireless drone communication for applications in environmental monitoring, surveillance, and research in dynamic indoor environments.
Video Link: – https://www.youtube.com/watch?v=XDKyIxvXuaQ
Demo-5:
Demo Title: Auto Switch between LTE/5G and WiFi with Prioritized Packet Delivery
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 102
Demo Exhibitor:
Jianbing Ni, Queen’s University
Demo Description:
We develop the following components to build the demo that achieves auto switch between LTE/5G network and WiFi network with prioritized packet delivery:
In Video 1: 1) A private network for LTE and 5G communication developed from the srsRAN software-defined radio. Mobile phones can successfully connect to the LTE/5G network through antennas; 2) a WiFi ad hoc network that enables the mobile phones to exchange messages within the network. The WiFi ad hoc network is built with the OpenWRT routers; 3) a LTE/5G signal detector that detects the LTE/5G signals; 4) a controller that achieves switch between the private LTE/5G network and the WiFi network with the raspberry pi. If the LTE/5G signal disappears, the WiFi network turns on and the mobile phones connect to the WiFi network, and if the LTE/5G signal appears again, the WiFi network turns off, and the mobile phones switch back to the LTE/5G network.
In Video 2: A packet delivery priority mechanism for LTE/5G network that enables the alert packets to be delivered immediately when the mobile phones are transmitting images with UDP.
In Video 3: A packet delivery priority mechanism for WiFi network that enables the alert packets to be delivered immediately when the mobile phones are transmitting images with UDP.
Video Link:
https://www.youtube.com/watch?v=pngx-mn2Lao
https://www.youtube.com/watch?v=wqXdcnp-lAA
https://www.youtube.com/watch?v=sJQznmRVw5g
Demo-6:
Demo Title: Human Activity Recognition using WiFi CSI
Exhibition Date:5 -8 Sept
Booth Location: Metropolitan East
Booth # 108
Demo Exhibitor:
Radomir Djogo, ECE University of Toronto
Navid Hasanzadeh, ECE University of Toronto
Hojjat Salehinejad, Mayo Clinic
Shahrokh Valaee, ECE University of Toronto
Demo Description:
The demonstration program consists of multiple Wi-Fi-enabled ESP32 microcontrollers, and an application on a computer with a GUI, enabling multiple key human activity recognition functionalities. Hand movements change CSI in the environment and is captured by the ESP32 microcontrollers. The real-time CSI signals is passed to a machine-learning model for gesture classification. The application is capable of real-time CSI monitoring from ESP32 microcontrollers, and can also record the real-time CSI, in order to use it for further model training tasks. The application can replay pre-recorded experiments, including video and CSI data, in order to demonstrate the functionality of the real-time HAR model in different environments. There will be multiple different pre-recorded experiments in different environments. The demonstration will also showcase different hardware that can be used for CSI data collection, including microcontrollers, Raspberry-Pi, and Asus Routers.
Video Link: https://drive.google.com/file/d/14XEjRI06N9vsm_xci4WtfvFSDseMkqct/view
Demo-7:
Demo Title: CottonCandy: A Low-Cost, Power-Efficient LoRa Mesh Network for Large-Scale Environmental Sensing
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 116
Demo Exhibitor:
Dixin Wu, Spero Analytics
Ahmed Mahmoud, Spero Analytics
Demo Description:
- a fully-functional CottonCandy sensory node, including the embedded system and hardware which run the network.
- A showcase of how the networking system works
- Pictures showing the sensory nodes and gateway we have set up at UofT’s testgrounds at Koffler Scientific Reserve (KSR)
- (If possible) we will show live environmental data from KSR being collected by the CottonCandy network installed there
Video Link: https://vimeo.com/851813007
Demo-8:
Demo Title: An Energy Harvesting Smart Bike
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth# 110
Demo Exhibitor: Nabin Chhetri, Mariam Mendha, Ahnaf Shahriar and Zihao Xie, Toronto Metropolitan University
Demo Description: The students have designed and implemented an advanced smart bike system that transforms a regular bike into a smart bike. This system monitors weather conditions, bike’s speed, inclination, location and the rider’s vital signs such as pulse rate while harvesting its own energy using a dynamo. The harvested energy is regulated and stored on a rechargeable battery so that the system continuously works even if the bike is not pedalled. The system also warns the user about approaching vehicles from behind using voice alerts. An app is also developed that enables hands free control of the system using a smartphone. The system measured parameters are also displayed and stored on the smartphone.
Video Link:
Demo-9:
Demo Title: Intelligent Systems for Active Noise Control within Aircraft Cabins
Exhibition Date:5 -8 Sept
Booth Location: Metropolitan East
Booth # 118
Demo Exhibitor:
Vishaal Venkatesh, Toronto Metropolitan University, Researcher
Reva Chaudhary, Toronto Metropolitan University, Student
Demo Description:
The experimental setup for the integrated system comprises three primary components: a speaker motion system featuring two speaker motors with encoders, a B1-E Dummy Head positioned on a turntable to capture head angular positions, and the Silentium control system. This control system takes noise generated by shakers within the cabin mock-up, applies necessary filters, and produces an inverted noise wave for recording average noise reduction. The demo video showcases the utilized speaker motion system during the experiments, along with the individual’s head exhibiting head rotations alongside the speaker motion system. The analysis of the integrated system provides a comprehensive examination of its components, encompassing the head tracking system, speaker motion system, and Silentium noise control system. These components collaborate to yield an intricate understanding of the integrated system’s diverse facets. The focal points of this study include the passenger head angle, the left and right motor angles and distance parameters . These geometric parameters are evaluated in relation to the Active Noise Reduction (ANR) variable derived from the Selenium system. The study assesses the impact of noise reduction by establishing various relationships, including head angle, motor angle and ANR.
Video Link: https://drive.google.com/file/d/1p4Q_1C1qB1ysQQHaH8ERRLGfenNWYxke/view?usp=sharing
Demo-10:
Demo Title: Demo of a Full-duplex link offering 115dB+ self-interference cancellation without compromising Links Noise figures – achieves higher cancellation if level of thermal noise permits
Exhibition Date: 5 -8 Sept
Booth Location: Metropolitan East
Booth # 122+124
Demo Exhibitor: Dr. Amir K. Khandani and Dr. Hojat Zamani
Demo Description: Performance of existing full-duplex links (to the best of our knowledge) are, without exception, expressed in terms of level of self-interference cancellation without paying attention to the increase in the noise level that comes as a byproduct. Various imperfections (incurred in the cancellation process, and/or in subsequent training synchronization/equalization) are packaged into a so-called Error Vector Magnitude. An increase in EVM is equivalent to an unfixable increase in noise level, which, at the end, increases the Noise Figure (NF). This degradation is typically several dB (10+ of dB), way above sensitivity level expected of modern systems. Indeed, since EVM is dependent on the actual signal, its damaging effect can be more severe than addition of a noise terms. EVM directly affects link stability and error performance. In addition, the added noise depends on the signal and the accompanying degradation in SNR cannot be compensated by increasing the signal power. This is unacceptable in modern wireless links where smallest degradations in NF, gap to Nyquist Limit and Gap to full utilization of power and bandwidth efficiencies are of utmost importance. Due to these shortcomings, worsened by various nonlinearity effects that surface prior to self-interference cancellation when the signal level is significantly below noise, application of full-duplex wireless has gradually lost its interest among industrial units. We hope this demo shows full-duplex wireless can function as good as or even better than legacy full-duplex links, and in turn contribute to rejuvenating the interest in the topic.
Demoed full-duplex modem relies on a hierarchy of tunable self-interference cancellation stages (both analog and digital), which, due to being a cascade of linear structures, can be measured/trained in succession. As a result, any potential error left in one stage of the hierarchy can be measured and corrected in subsequent stages of hierarchical echo (self-interference) cancellation. Another key distinction with previously reported full duplex links lies in the introduction/incorporation of pairwise symmetrical antennas which play a decisive role in reducing self-interference in the analog domain and thereby allows the digital cancellation to operate at its required reduced self interference level. For the first time, our setup fully relies on linear components as well as linear operations, utilizes off-the-shelf components that are readily available. Components such as ADC/DAC, Crystals, etc. that affect the performance and accuracy are selected among entry level components in the market, all enjoying wide range availably and very low cost.
Each setup relies on a single full-duplex antenna at each end (in most cases each full-duplex antenna is a fully connected piece of metal) and isolation between transmit end and receive portions are caused by novel geometrical symmetries and adaptive analog filter(s) that cause cancellation of unwanted signals and thereby allows the digital cancellation to operate at its required reduced self interference level.
This demo will demonstrate real-time over-the-air transmission of two parallel wireless links supported by a single full-duplex antenna, while occupying the same spectrum. The setup includes fully functional modulation (OFDM), coding (convolutional code), modulation (QAM over each of OFDM tones), specialized full-duplex OFDM training (for self-interference cancellation as well as receiver synchronization). In a nutshell, it offers: (1)-Full utilization of available bandwidth (realizing Nyquist limit), simultaneously and in both directions. (2)-Achieving a signal-to-noise ratio, simultaneously and in either directions governed by thermal noise level, i.e., with a truly negligible Error Vector Magnitude added.
Several full-duplex, wide-band antenna structures suitable for applications such as: (1) LTE/5G/6G (2) Wi-Fi, (3) White Spaces, etc. will be presented. In addition, role of full-duplex wireless in simplifying point-to-point massive MIMO, in new collaborative signaling schemes that would not be possible otherwise, in military operation compromising self-protection through emission of controlled emission of multi-user interference (controlled jamming), and significant facilitation in space-division multiple access will be discussed.
Note: Single links forming the full-duplex connection rely on a channel coding and modulation, OFDM structure, tracking and equalization, and other designs and data recovery operations like what is deployed in conventional 802.11G, but of course with the exception of formation and management of superimposed full-duplex links.
