2015 Award Winners
Grand Prize Winner
Worcester Polytechnic Institute
Search and rescue robots have been in use since the September 11, 2001 attacks on the World Trade Center. They were deployed to search for victims, identify hazardous materials, and provide human rescue crews with a thorough understanding of the dangerous environment. Robots were also deployed after natural disasters including hurricanes Katrina, Ike, and Sandy, and the 2011 tsunami in Japan. These situations presented an additional challenge for rescue teams because of pervasive flooding, which could not be overcome by commercially available land rovers.
Our proposed solution is the Water and Land Remote Unmanned Search Rover (WALRUS Rover), an amphibious rover to aid in the search and discovery of survivors. It will feature shared autonomy, high-definition vision systems, payload capability, two-way communication, and advanced mobility. Such a rover would not only give relief teams more information about the situation at hand, but also eliminate the danger of sending search parties into harsh and high-risk environments. The WALRUS Rover will be capable of overcoming indoor and outdoor obstacles like stairs and rubble as well as navigating in flooded environments, enabling a wider range of mission profiles.
University of Pennsylvania
Our team has found the next area of improvement for prostheses that can increase the functionality at a reasonable cost: add temperature, pressure, and orientation feedback to the user.
Two million people live with limb loss in the United States, and hundreds of thousands of United States citizens are added to this statistic each year.1 Creating a prosthesis that functions intuitively for these amputees is imperative for two reasons. First, using a prosthesis that has added functionality will translate to helping the amputee perform day-to-day tasks. Second, succeeding in creating an intuitive prosthesis is imperative because of the impact of phantom limb pain (PLP). PLP is described by aching or other uncomfortable sensations perceived at the site of the amputation. A prosthesis which correlates body signals to its movements may eliminate PLP and thus improve the user’s quality of life.2 Adding feedback will increase this correlation, making the prosthesis more able to benefit the lives of millions. As a result, our team will be developing a low-cost, 3D-printed, myoelectric-controlled, haptic instrumented prosthetic arm for trans-humeral amputees. The system will also include haptic presentation of the tactile data collected, and we plan to make it an open-source project.
Pennsylvania State University
Our goal is to create a add-on device for 3d printers that can turn any 3d printer into an automated device. This device will incorporate several features like monitoring and wifi to provide the users with the ability to send prints to a queue to be printed. This device would be the centerpiece to our companies 3D printer automation device, which allows users to print multiple objects in a row.
University of Pennsylvania
DORA is an experience-driven, entertainment-based navigation platform that will allow users to conveniently, economically, and thoroughly explore remote locations. We seek to capture the enthrallment, inspiration, and sense of belonging associated with traveling to lands both familiar and foreign. As such, there will be two main thrusts: (1) the “end-user” experience, equipping the average consumer with the necessary sensory stimulations to feel fully immersed in the environment that he is exploring, and (2) the exploration platform itself, which will ideally be capable of traversing extreme arenas inaccessible to casual visitors, or, by nature of their environment, otherwise preventing visitation.
Soft Robotic Hand
Worcester Polytechnic Institute
Conventional hard manipulators require a high degree of precision when attempting to grasp an object, and typically are limited to a small range of tasks. Robots capable of high precision manipulation are prohibitively expensive for many applications. Lower-cost robots, such as Rethink Robotics’ Baxter, lack the precision to interact with small or delicate objects.
Our solution to this problem is to design and develop a soft robotic manipulator that can lessen the precision required with robotic grippers today. The design will incorporate fluidic elastic actuators constructed out of silicone rubber, and will have soft, deformable sensors embedded within the actuators for position and force feedback. The feedback systems will allow the force output of each actuator to be controlled individually.
By focusing on safe, reliable grasps with limited precision, our gripper is able to manipulate objects that would otherwise require a much more expensive system. The gripper is projected to be low-cost compared to other hard manipulators. Existing soft robotic grippers, such as the RBO hand and PneuNet technology lack controls necessary to safely interact with objects requiring precise movements.
University of Pennsylvania
In the automotive racing industry, the most commonly used head restraint system is the HANS device. This device was developed in the early 1980s with the goal of decreasing the risk of basilar skull fracture during automotive collisions. It uses two rigid straps to constrain head motion during a crash. Drivers using this restraint system commonly complain of feeling restricted by the device, and while the risk of basilar skull fracture is much lower, concussion and other mild traumatic brain injury are still common.
Our team will design, build and validate a head and neck support system for automotive racing that uses active dampers for energy dissipation. The product will improve on current options by decreasing the rates of traumatic brain injury during collision, by increasing driver visibility and mobility, and by lowering the chance of peripheral injury such as collarbone breakage. An active damper system will allow for low speed movement to be unaffected, but precise and tunable control over head deceleration during a crash event.
Our group is determined to make a user friendly, automated hydroponic growing system for people to grow food and have fun doing so. The user will not be required to have any prior knowledge of growing plants, or understanding of what is happening behind the scenes. GrowBox will be where the plant resides and what sees after its needs. Behind the scenes there will be a system of sensors and image processing software to provide data on the plant’s status. An iOS app will act as an interface for this smart appliance, notifying the user when something is required of them, and what is required. GrowBox is intended to bring fresh, home-grown vegetables to those who don’t have the time, space, or experience to grow their own.
University of Colorado Denver
The field of computer vision seeks to develop machines with the capabilities to perform tasks that currently can only be performed by humans. Important aspects of human vision such as the ability to identify obstacles, plan paths around them, learn from one’s surroundings, in real time, area crucial to many societal uses of technology. Each year, computer systems running complex vision software collectively step closer to emulating each stage of human vision. However, accurate and efficient computer systems have always been obstacles to integrating greater amounts of automated robotics to help people. To enable next-generation robotics to perform field sensing in dangerous and remote environments, Stratus will deploy a sophisticated real-time Simultaneous Location and Mapping (SLAM) algorithm to provide precise odometry, high-resolution 3D environmental maps, and awareness of specific adverse environments. The Stratus team will investigate the use of 3D sensing and computer vision technology available for mobile platforms and design custom algorithms that trade-off execution time, accuracy, and power efficiency. Overall, the goal is to demonstrate a large-scale near-field mobile 3D mapping system to aid in rapid disaster recovery and environmental analysis scenarios. Specifically, Stratus will aid in surveying Colorado Rocky Mountain rock slides and avalanche areas.
SCRAM (Supply and Command Rover for Autonomous Multicopter)
University of Central Florida
This project will see to the design and creation of an autonomous rover whose main purpose will be to communicate with, and charge, a semi-autonomous multicopter. The rover will be able to wirelessly issue way-point commands as well as designate preset paths for the copter to follow. These preset paths will be uploaded to the rover from an outside station using long distance communications and then relayed to the multicopter using short distance communications. When the copter’s battery life has dropped to a certain percentage the rover will be notified via radio transmission and a rendezvous location will be set by the rover based on the relative locations of the two vehicles. When the copter and rover come within close proximity to one another, a targeted autonomous landing sequence will commence. Once the copter has safely landed on the rover’s charging pad, the solar energy which has been gathered will proceed to charge the copter’s battery. While charging, the copter will wirelessly transfer the sensor data accumulated during flight to the rover, subsequently freeing up storage space for its next flight. Once the charging and data transfer is completed, copter and rover will separate and the mission will be resumed.
The number of bicycle accidents on the road increases every year. While bicyclists are expected to take extra precautions, motorists can also help prevent bicycle accidents with the assistance of a bike detection system that helps improve a driver’s road awareness by alerting them to potential crashes before they happen.
The C.A.R.R. System (Cyclist Alert Real-time Response) is an all-in-one alert system which checks both sides and behind your vehicle for approaching cyclists. Using a sophisticated real-time detection algorithm, The C.A.R.R. System instantly notifies you of potential and imminent collisions to assist your maneuvering before a crash occurs.
The system consists of an easy to install mountable alerting system and features an intuitive driver alert interface. The system uses two cameras, one on each side-view mirror, which feed an image-processing algorithm. If a cyclist is positively identified, the alert hub will generate a non-distracting audio and visual feedback on the central dashboard informing the driver of a danger and from which side this potentially dangerous cyclist is approaching.
Seattle Pacific University
CERBERUS is an autonomous fire-fighting robot to be stationed in buildings, capable of first-response fire suppression without any human intervention.
University of Massachusetts, Lowell
Traffic congestion is a condition where slow moving vehicles congregate, causing increased travel time and fuel consumption. The average American wastes 40 hours and $818 a year due to traffic congestion. Collectively, the American population will spend around $121 billion annually on fuel consumption due to high density traffic. By observing historical trends in traffic data, it is possible to mitigate the adverse effects of traffic congestion by improving the infrastructure for highways, subways, and other forms of public transportation.
Current systems used for data acquisition can be intrusive and some are prone to error depending on weather. One recent solution is the use of Bluetooth sensors. With the increasing popularity of mobile devices, a chain of Bluetooth sensors can be used to analyze traffic data based on the location of a passenger’s Bluetooth enabled device. However, the amount of people who constantly have Bluetooth enabled on their device is low.
NetLane aims to increase the sample size of data acquisition by using Wi-Fi sensors. Initially, NetLane will target buses to determine the volume of passengers at each stop as well as their origin and destination. Our solution will provide a scalable method of non-intrusive data collection for transportation engineers.
Arizona State University
When the landscape is torn up by natural or man-made disasters, location of victims, deployment of resources such as food, water and medical supplies and establishing communication becomes an enormous challenge. Even if we have the required resources, all of the effort goes in vain if we cannot deliver them to the victims. Therefore, identification and location of victims, at the time of a disaster is of primary importance for any relief worker. Speedy detection allows the relief worker to save precious time, which is critical for preventing and minimizing casualties.
We propose an Airborne Disaster Relief Assistant (ADRA), which will help in locating a victim, provide GPS location to the personnel for quick response and deliver important packages such as first-aid kits, food rations and communication aides to the victim. The ADRA will have optical recognition and heat sensing capabilities for human detection purposes, and the data collected can be transmitted to a base camp for analysis and decision-making. It shall also possess an attitude adjustment system for maneuvering and collision avoidance.
In order to educate a patient under physical therapy, a physical therapist usually needs to visually identify what physical problems the patient is facing and whether the patient is practicing correctly. It is useful to video-record a patient’s movements when he or she is practicing at home or during a distant physical therapy session. Currently there is no convenient way of fully recording a patient’s movements.
Therefore, our team proposed the design of “Guardian Angel”, which is a tracking system docked on an unmanned aerial vehicle (UAV) that carries a camera to video-record a target’s movements. The user will be able to control the position of the UAV upon initialization and let the UAV follow it’s target. Our design consists of two major parts: user-end devices and UAV-end devices. On the user-end, there are two devices, including an Android device programmed to control the UAV, and a helmet equipped with a position-transmitting beacon. On the UAV-end, there is an Intel Edison board serving as the central computation unit and it communicates with a flight controller to control the UAV movements. We are using APM 2.5 flight controller to stabilize and control the flight, and we are using the triangulation positioning algorithm as well as computer vision algorithms to implement tracking and following.
University of Pennsylvania
BAM! 3D or BAlloon Mason 3D is a 3D printer that is supported by a balloon. The balloon will be attached to cables that are loaded in tension. The length of each cable is controlled by a ground mounted motor, which will allow the printer to move to any specified location. This project will allow structures to be printed without having a frame around the object, which will be beneficial for cost effectively printing buildings and other large objects. It will also be a compact solution to 3D printing buildings, which will allow it to be easily deployed to many locations. Our project will build a smaller scale model that will be able to print with concrete or a similar material.
The Tufts Robotics Club plans to use an Intel Galileo in the construction of a universal gripper arm. The universal gripper, a novel gripper technology developed by Cornell University, utilizes the jamming of fine material to pick up objects. A balloon filled with fine material is pressed onto an object. Air is sucked out of the balloon causing the material inside to constrict around the object, allowing it to be manipulated. Here is a video: http://www.youtube.com/watch?v=0d4f8fEysf8
The arm portion consists of two joints and a base able to rotate 360 degrees in order to achieve full range of motion. A challenge in building the arm is accounting for its weight. The arm will use timing belts attached to servos located at the base to generate a very accurate and quick moving arm. The timing belts will also offload the weight of the motors from the arm which will add more lifting capacity. The arm lengths will be constructed from PVC to make the arm even more lightweight.
The arm has applications in mobile robots tasked with manipulating their environment. Additionally, this type of arm mechanism is extremely flexible in that it is able to pick up virtually any kind of object.
University of Rochester
NullSpace VR is a multisystem computational and hardware platform for environmental virtual reality, capable of tracking a user’s movements, translating them into a virtually constructed space, and providing full body haptic feedback when the user comes into contact with a virtual object. In layman’s terms, this allows a user to “feel” a virtual object simulated in front of them via a computer program, in conjunction with visual feedback with a VR headset such as the Oculus Rift. The project is currently capable of performing this task using vibrational feedback on the hands within a confined space, using infrared camera tracking to interpret the user’s location. Our future goals for the project include a hybridized system for tracking that fuses 9 axis sensors with camera motion capture, multi-camera tracking for large and flexible environments, and more sophisticated methods of force feedback for user interactions.
University of Pittsburgh
WASP is a small unmanned aerial system (SUAS) able to track and follow multiple GPS transponders. Its purpose is to improve current SUAS systems employed by the military and present a new method for employment in the field. Current systems lack autonomy, have excessive training requirements, and require a special team to be utilized. WASP will address these issues to make the system easier and more effective to use.
Elephant Tracking Device
University of Akron
Elephant endangerment is a continuing problem around the globe with the World Wildlife Federation listing species of elephants as three vulnerable, five endangered and one critically endangered. Current conservation methods include tracking systems installed as a collar worn by the elephant. Researchers have also made advancements in the understanding of the infrasonic communication used by elephants across long distances. The current collar design has a finite lifespan and requires sedation of the elephant to install and repair. Employing wireless power charging through magnetic resonant coupling increases the amount of power available on the collar while also extending the lifespan of the device, allowing higher resolution tracking and the introduction of an infrasonic audio system. The goal of the proposed elephant monitoring device is to provide real-time access to infrasonic elephant communication and location while extending the lifetime of the collar. Acquisition of infrasonic sound data from the vicinity of the wild elephant in conjunction with the elephant’s location will allow researchers to study this unique communication method simultaneously with elephant behavior in their natural environment. Increased field life of the device also requires less maintenance and reduces human interference with the animal.
University of Scranton
The invention is a motion controller generally capable of operating a moveable device in multiple dimensions/axes to track an object. Example devices controlled in proof of concept design include an astronomical telescope and reflective solar tracker array. For the reflective solar tracker, the motion controller utilizes sun astronomical data from the U.S. naval observatory to continuously keep solar panels sun facing maximizing their energy output. Future embodiments of the motion controller include user inputted GPS coordinates to point a device at the desired location.
Oregon State University
Our project entails the creation of a wireless network that communicates between vehicles and an interface within the vehicle to create an overall safer driving experience. Our challenge is to make drivers more aware of their surroundings. This challenge can be overcome by creating a family of embedded devices that can communicate with one another in order to notify drivers of potential hazardous situations that the devices and algorithms detect. This will make roads a safer place by giving drivers real-time information updates from other vehicles in the network. These warnings and messages will be displayed on an interface inside the car. Also, upon the interface there will be the car's information displayed, that isn't normally available to the driver.
Ever wonder what it is like to communicate without speaking? You can try using hand gestures or home signs, but the problem is that they are not universal. You need to learn a sign language that everyone understands, but unfortunately there doesn’t exist any. Even a widely accepted language like the American Sign Language (ASL) is barely understood by anyone outside the hearing impaired community. This is a problem all the hearing-impaired people face in their everyday life.
Our Project slate8 would help hearing-impaired people communicate conveniently with people unfamiliar with American Sign Language (ASL), and hence bridge the communication gap. This will also raise their self-confidence as they can better adapt into the broader community. The system will employ Intel atom board with a camera to stream video as input, which is then converted into text/voice using real-time image processing algorithms.
Intel-Cornell Cup Exhibition Team
Federal University of Pernambuco
The everyday task of navigating one’s surrounding can amount to a true challenge to the visually impaired. Buildings are usually not designed with the challenges of this particular group of people in mind, which further aggravates the issue. Guide dogs have long been the most common attempt at solving this problem. However, guide dogs are quite limited, for they are incapable of reading signs and other visual clues, which humans use naturally, and they cannot perform navigation. Furthermore, the cost and effort involved in acquiring and caring for a guide dog may be prohibitive to some. We aim to produce a system that is both lower in cost and richer in capabilities than the traditional solution to the problem, namely, guide dogs, while also being easier to obtain. We propose our solution in the form a robotic car capable of guiding its user around buildings. The car is equipped with sensors that allow it to follow a path, avoid obstacles and determine when it has reached a destination. The system will be also capable of taking in voice commands, which in turn removes another barrier to the visually impaired.