Bio-sensors are becoming an integral part of our everyday life, implicitly and explicitly. For e.g., health monitors, food quality monitoring. The natural sensing mechanism and advancements in synthetic biology that allows to engineer DNA to alter specific traits of bacteria has made bacteria as potential candidates for building bio-sensors. The ubiquity of bacteria also facilitates bio-compatibility. With advancements in synthetic biology, several synthetically engineered circuits have been developed for oscillators, switches, sensors, and storage. The successful operation and availability of biological circuits and components for data processing make bacteria strong candidates to use as computing machines. Genetically engineered bacteria can, therefore, be used to sense and perform computations on the data sensed.
Currently, bio-sensors are used purely for sensing and the information is processed off-line. This leads to further delays and manual post-processing. We propose to build a network of bio-sensors that can communicate with each other in real-time and provide a continuous monitoring autonomously. Such a network of bacterial(bio) sensors differs significantly from traditional electromagnetic communication due to the devices used, channel, environment and the application.
We use Escherichia coli (E. coli) bacteria genetically engineered to exhibit fluorescence upon the receipt of a specific signal molecule (N-(3Oxyhexanoyl)-L-homoserine lactone, or C6-HSL). A microfluidic experimental system houses bacterial populations within mmicrometer-sized chambers fed by channels that provide both nutrients and controllable levels of C6-HSL, to demonstrate that a chemical signal at the sender can be reproduced as a fluorescence signal at the receiver reliably.
Time Elapsed Communication:
We demonstrate that it is indeed feasible to realize a simple modulation technique such as On-Off Keying (OOK) for communication between the bacterial populations, but the consequent data rates achievable is as low as 10−5 bps due to very high processing delays at the receiver. We term such environments where the processing delays are very high as Super Slow Networks. Existing modulation techniques are energy based and the number of signals transmitted is proportional to the message length.
We introduce a communication strategy called time-elapse communication (TEC), wherein information is encoded in the time period between two consecutive signals. The number of clock signals elapsed between a start and stop signal encodes the information to be transferred. Thus, the number of molecular signals generated always remains at two (the start and the stop) irrespective of the number of bits required to represent the information. Intuitively, TEC improves the data-rate over OOK by reducing the number of communication signals that need to be conveyed per unit of information.
Amplitude Division Multiple Access:
A star topology with multiple sources and a single receiver can be most commonly observed in sensor networks, where multiple sensors communicate to a single receiver/sink. When multiple sensors in a network report to the receiver, they broadcast information to the receiver and thus do not require the destination address. The receiver, on the other hand, receives signal from multiple sources and hence needs to know the address of each source. We define such an addressing mechanism as Source Addressing. In a super-slow network like bacterial communication, address fields and bits increase the overall redundancy. We propose a source addressing mechanism called Amplitude-Division Multiple Access/Addressing (ADMA) , which uses the amplitude of the signal transmitted as the address of source. Each source is assigned a unique amplitude. The sources transmit rectangular pulses with the assigned amplitude. When multiple sources transmit simultaneously, the receiver receives the sum of amplitudes, from which the receiver identifies the components of the sum. Thus, the address of the sources implicitly solves multiple access control.
Interference Management In Distributed WiFi Networks
In a managed WiFi network, radio resource management (RRM) is performed with the help of WLAN controller that acts as a central controller. In public hotspots and residential deployments, the APs are independent of each other and contend without the knowledge of other overlapping BSSs. In such distributed network, there is need for a central controller but none of the independent APs have the motivation to have a central controller.
We developed a parliamentary approach to achieve cooperative allocation. An AP is chosen as a local group leader based on 2-hop network information. The chosen leader assigns channel to its neighbors. The leader knows its 2-hop neighbor and develops a centralized decision for its neighbors. The decision to be made can be resource allocation, scheduling etc. This mechanism is repeated periodically to approach the performance of a managed network. The grouping mechanism achieves the performance of an enterprise network with WLAN controller without using an explicit controller. We compare the number of OBSS per AP when using QLoad report and grouping mechanism and show that the grouping mechanism reduces the number of OBSS.
A shift towards wireless
With the ubiquitous presence of WiFi, for a cable ISP, the end-user experience is determined by the wireless performance. Therefore, the customer satisfaction of the cable network is determined by the efficiency of the wireless network.
In an unmanaged network like residential deployments, inter-cell interference triggers backoff mechanism of CSMA/CA and further affects the latency. We proposed a new architecture that reduces cable deployment and replaces cable with wireless without compromising on the end-user experience. The last leg of the wired network is the Cable Drop, that connects the end-user homes to the cable system. We propose Wireless Drop, a wireless substitute for cable drop. Wireless Drop saves cost on laying out cable for new deployments and makes it easier to fix or repair in the drop. A wireless drop can serve more than one home, thus creating a centralized architecture to perform radio resource management. We designed and performed extensive test cases to determine the efficiency and effectiveness of wireless drop in a real-time deployment.