Assignment 4: Parts 1 – 4
Part 1
To find the original project proposal, please visit my Assignment 4: Project Proposal webpage.
Part 2
a.
To find the live data of the University of Oregon’s inductive loop bicycle counter, please visit my Assignment 4: Part 2A page.
Our first graph displays how bicycle count averages differ from the university’s enrollment data. This visualization is integral to our study, as it contrasts the total amount of commuters (or a number close to the actual amount, assuming most students attend the classes they are enrolled in) to those choosing to ride their bikes and continuing past the installed bicycle sensor on 13th and Kincaid. The graph on the left shows us the hourly rate at which cyclists arrive at campus as opposed to the amount of students who are (expected to be) on campus. Specifically, it shows that cyclist counts peak each day before and after peak enrollment times. Additionally, the largest peak of bicycle commuters occurs in the morning, allowing us to assume that bicycle riders arrive on campus in the morning, and leave at various times thereon after.
The second graph reveals bicycle count and enrollment data to be compared with average temperature and rain on a daily basis. The graph reveals that aggregate bicycle counts align with enrollment data, and that bicycle counts were not as high when either the temperature or precipitation amount rose. This simplistic visualization was chosen so that all factors could be compared with each other at once without risking a cluttered or confusing interface.
The sensor used in the automatic bicycle counter is an inductive loop counter, which is considered to be the most reliable traffic detection method available on the market today. It consists of mechanisms installed underground that produce small inductive loops detecting the presents of metal. These loops are constantly active, and when a bicycle tire passes over the mechanisms, the inductive loop is interrupted, and thus a bicycle is sensed. As the inductive loop counter is collecting data, transmitters from the sensor nodes send the data to a receiver and is distributed and stored via ethernet. Once the data has been sent to the ethernet it is available for collection. We receive the bike counter data in an application programming interface which has been special designed to format the data into either a GeoJSON or CVS format. This is considered to be an open sensor network since this data may be remotely accessed and altered at any time.
b.
The visualizations above once again aim to show how campus enrollment data contrasts with bicycle count data. However, this time we gather the bicycle count data from the CLMPO’s Eco-Count sensor dataset which takes climatic variables into account as well. The visualizations were chosen to emphasize population count differences between the CLMPO data and climatic variables as well as campus enrollment data.
The leftmost graph allows us to contrast the CLMPO bicycle count data with the university’s enrollment data. While paying close attention to the dual axis (as they are scaled differently), we observe that the largest disparity between these datasets occurs on the weekends where bicycle count data differs largely with respect to enrollment data, though it is lower than the aggregate values of other weekdays. Thus, we can conclude that enrollment data has a loose correlation with bicycle count data; it is generally closely related to bicycle transportation with the exception of weekends.
Because the CLMPO data has been collected over a longer period of time than the UO’s automatic bicycle counter, I take advantage of this in our visualizations, and choose to visualize bike count and enrollment data with two different climatic variables on a monthly scale. These graphs reveal that bicycle counts lower when precipitation increases, and when minimum daily temperature decreases. As expected, enrollment data fluctuates independently from both climatic variables.
The CLMPO has used a portable bicycle counter called the Eco-Counter’s Tube sensor. This sensor consists of two pneumatic pressure sensors encased in rubber tubes. The tubes are meant to be oriented perpendicular to traffic on a road. When a vehicle (a bicycle or car) passes over the rubber tubes, the phenomena of the pressure exerted on the tubes is sensed. Between the two tubes, the distance between the front and back tires of the vehicle are sensed. This information, along with the speed of the traveling vehicle are included as parameters in an internal algorithm that deduce whether or not he vehicle was a bicycle, or some other object.
The Eco-Counter capitalizes on the sensor network’s ability to allow data to be visualized on mobile applications. The sensing system comes with GSP (Global System for Mobile Communications) transmitting ability, and thus allows data to be easily visualized on a mobile device (this is at least true for the newest version of the counter). Additionally, the bike counter contains 2-years worth of memory to allow the bicycle counter data to be stored and accessed manually. In this case, the Eco-Counter sensor network may be used as either a closed or open sensor network, as the data must by manually extracted, or can be visualized live with a mobile device.
Part 3
Integrating Bicycle and Public Transportation in Eugene and Springfield
Introduction
For decades transportation planners have been plagued by the concept of ‘the last mile,’ a truism that represents the difficulty of transporting people from major transportation hubs to their respective destinations. With respect to the growing popularity of cycling as a major mode of transportation in Eugene and Springfield, this problem has a near perfect solution: an integration of bicycle and public transportation. Easing the ability for bicyclists to incorporate public transportation into their commute (and vise-versa) can allow citizens to experience a reliable, accessible and sustainable commute.
The aim of this project is to utilize sensor and wireless sensor network technology to provide citizens with foresight to judge whether or not they will be able to transport their bicycle on a given EmX bus. In pursuing this focus, I have designed a system which will sense the amount of bicycles being carried on a single bus at any given time. This will be done through the implementation of an open, wireless sensor network which will make use of strain gauge load cells to sense the weight of a bicycle. These load cells will be situated directly underneath the bicycle racks on each EmX bus, and will sense the existence and weight of an object situated on top of the sensor. The sensed data will then be transmitted by a transceiver to an external data gathering devices which may advance the data to other routers in the network. This data will be visualized on simple LED monitors which will allow the information to be clearly displayed and available for the public to view.
To elaborate on specific operations and entities within this system, I will address the following lines of inquiry to guide the explanation of my design: (1). What specific technologies will be implemented in this network? (2). How will these devices interact with one another to receive and deliver the desired information? (3). How will this information be delivered to its recipients? (4). What implications might this system have on city-scale transportation? (5). What are ways to possibly improve this design? (6). Are there ethical or privacy violations that should be acknowledged in the system>
1. Technologies Overview:
This system will reflect that of a typical open wireless sensor network. Thus, the network will make use of individual sensor nodes (which will be placed on each EmX bus), multiple sink nodes and visualization screens (to be placed at each EmX bus stop in the Eugene-Springfield EmX corridor).
The sensor nodes will each contain the following devices:
microcontroller – a very small, battery-powered, autonomous device that is capable of executing simple algorithms, and storing memory locally.
strain gauge load cell – a sensor shaped like a cantilever with an attached resistor which senses the distortion of the cantilever arm’s material when force is applied. These sensors are commonly used to sense weight accurately. This particular model is able to weigh objects accurately up to 50 kg. Since the weight of a stationary bicycle is what we are aiming to measure, we choose a strain gauge load cell to measure the weight of a stationary bike to determine whether or not a bike is present in a bike parking area on an EmX bus. We use weight to determine bike existence because it is not affected by the temperature or color of the bicycle (which may pose a problem if we decide to use an infrared motion sensor). Sensing weight with a strain gauge load cell, as opposed to a force-sensitive resistor, allows us to accurately determine the front end of the bicycle’s weight so that we can sift out force values that are less than the lightest front end of a bicycle (since bicycles may carry cargo, we are more flexible with receiving higher than normal weights).
short-distance wireless transceiver module – a transmitter / receiver that is capable of both transmitting and receiving small data packages via wireless transmission. The distance at which data may be transmitted is adjustable, but for the purposes of our project lies around 10 – 30 meters.
Other devices in the system include:
router outfitted with 802.15.4 Zigbee – this specific router is weather-resistant and may receive data transmission from a sensor node, and allows the data package produced by the microcontroller to be forwarded to other routers in the network.
LCD visualization screen – the simple, durable LED screen allows easy, low-power visualizations to be displayed.
Estimated Cost of Technologies:
2. Building an open sensor network:
Each sensor node will consist of three strain gauge load cells, a battery pack, and a transceiver which will be connected to the microcontroller. Once the strain guage load cell has been stimulated, an analogue to digital converter will convert the sensation to voltage. This volatage is forwarded to the microcontroller. This device will be programed with a simple internal algorithm that makes it responsible for the simple task of determining weights that are too insignificant to be a bicycle (some value under seven pounds for the front end of a bicycle, this value includes even the lightest bicycle designs).When the weight of a bicycle is sensed, the signal is then transmitted to the next node in the system via wireless transmission, this node being the sink node, in our case, a router.
The router will be fixated in some secure location of each EmX station, and will be equipped with the 802.15.4 Zigbee – a specification for high-level communication protocols that is composed of multiple low-powered digital radios. Once information is received by the data gathering system, it can be relayed to other routers in the system through a mesh network, a type of network topology that allows data to be passed through a series of intermediate devices to an end destination, which we will define as being the following bus stops in the same direction as the respective bus. When the data is received through the router with the Zigbee module it will activate a simple visual monitor that will display how many bicycles are currently occupying on-bus bicycle racks since the last stop, and how many spaces are currently open for incoming bicycles.
3. Communicating the Data:
The collected data will be displayed on a simple led screen at each EmX bus station. Three lines will represent the three respective bicycle parking spaces on the incoming EmX bus. If only a line is displayed, this will signify that a specific space is open on the incoming bus. Conversely, if a small bicycle symbol is situated above the line,this will signify that a bike is currently situated in one of the parking spaces (prototype shown below, where two spots are occupied, and one is availible).
Because this data is real-time and – in the scope of this system – is only useful for a few minutes, this will be the only visualization needed to communicate the desired information to its recipients.
4. Expected Impacts of the data:
The dataset this system creates will be highly volatile. Because the data is both captured and used within minutes, the data loses its utility quickly and is not stored. This means that no archive is kept to signify the usage of bicycle racks and the volume of the dataset is kept relatively small. Though the dataset this system produces is small, and certainly not ‘big data,’ it provides an extremely useful tool for citizens due to its volatility.
One of Eugene’s unique features and largest challenges is its suburban sprawl – the suburbs around Eugene extend out for miles, causing large populations of people to reside far away from areas of interest. This makes transit network design a challenge, as it is both expensive and inefficient to cast bus systems to far-reaching areas where only a few citizens may travel, and equally so to keep these populations from riding the bus. To retain the equitable mobility public transportation strives to achieve, there is a certain balance that must be struck between how far residents are expected to transport themselves, and where public transit picks up. This dataset is thus able to increase the fluidity of a multi-vehicle commute by providing volatile, real-time data about one of their chosen modes of transportation that will allow their commutes to be more convenient. This means that the value and validity of our dataset is mainly subject to its volatility.
In general, this system would be implemented with the hopes of increasing both bicycle ridership and EmX ridership to create a more livable, healthy, sustainable city.
5. Ways to improve: Non-Technological Design Proposals:
Because the data produced in this system is real-time, and meant to aid commuters in their daily decisions, it is integral to this system that the dataset be as accurate as possible. To improve upon the sensor network, a number of non-technological design choices are proposed. These proposals include the following: (1). Retrofit on-bus bicycle racks to accommodate the strain gauge load sensors, (2). use mechanical linkage to center the applied force on same area of the sensor during each usage.
Bike rack design and infrastructure must be retrofitted to increase the accuracy of collected data. This should be done by placing the sensor in a trough wide enough for only a bicycle tire to settle in. This will further ensure that the weight of the bicycle tire is on the correct part of the strain gauge load cell. Because this data is real-time, and is displayed directly to EmX riders, the data must be as accurate as it can possibly be. Thus, the retrofitting of bicycle racks is a necessary step in the implementation of this sensor system.
Due to the strain gauge load cell design, it should be noted that in order to receive an accurate weight, force must be applied to the same part of the cantilever arm at each use. This variability of force placement will be restrained by encasing the sensor in a hollow, plastic box containing mechanical linkage that directs force onto a single consistent part of the cantilever. This design would be very similar to that of a digital scale.
6. Ethics and Considerations:
The speed at which this data is discarded allows it to avoid the majority of the scrutiny big-data producers and smart-city technologies receive. However, there is a smaller considerations that must be acknowledged.
One safety concern that may arise out of the displayed data is the possibility of bike theft on a crowded bus. Displaying the location of an unlocked bike is always a danger, and the fact that bicycles are not locked to the bike racks inside of EmX busses may raise the chances of bicycle theft.