Swarm Capable Autonomous Vehicles (SCAV)                       

A test-bed of networked micro-robotic vehicles has been developed in our laboratory for carrying out mobile Ad Hoc and sensor network protocol research. As shown in the picture below, the SCAV platform has been developed as a 6"x6"x5" micro-robotic sensor system with the following capabilities:

•  Self-localization at centimeter resolution within an in-laboratory reference coordinate system

•  Both Ad Hoc and infrastructure based wireless communication 

•  Wireless networking using CSMA and AODV as the MAC and the routing layer protocols

•  Autonomous navigation using a second order Kalman Filter and the localization system

•  Collision avoidance using infra-red obstruction sensing

•  Multi-modal onboard sensing including acceleration, temperature, magnetic field, light, and sound

•  Remote debugging, command control, visualization and software download over wireless

Integrated Testbed: The developed mobile sensor testbed and its integration with the existing wireless network infrastructure are shown below. The testbed contains the following main components: 1) a fleet of indoor mobile SCAVs, 2) an indoor localization system for enabling sensor self-localization in GPS-denied environments, 3) a control, debugging and management infrastructure, and 4) a tiered wireless ad hoc network for seamless integration of the above three components and the existing research infrastructure in various wireless network laboratories in Michigan State University.

Control Debug and Management System (CDMS): The CDMS software system is designed for managing the SCAV infrastructure locally or through a network as shown in the diagram abobe. The CDMS system runs on a Windows system and supports the following essential operations for the SCAV testbed: 1) compiled image and software download using wireless links, 2) SCAV system debugging by allowing remote print console, 3) command line and graphical interfaces for remote command dispatch to targeted mobile sensors, 4) supporting a variety of sensor and mobile ad hoc network protocols for seamless connectivity with the SCAV testbed and the backbone mesh network, 5) real-time location tracking for individual sensors and their experimental post-processing,  6) command driven SCAV navigation, 7) sensor specific mobility planning through a graphical user interface, and 8) data upload and telemetry from the SCAVs using wireless links. Real-time location tracking screenshot from an example 2-SCAV leader-following experiment is shown in the following diagram.

Demo Video: This video demonstrates the SCAV system and its applications including collaborative mobile sensing, networked target tracking, and Kalman filtered navigation. The video also shows real-time CDMS view for all the applications.

Ongoing Research: The SCAV test-bed is currently being used for the following research activities:

•  Impacts of controlled mobility on wireless network performance

•  Collaborative networked navigation for multiple agents

•  Integrated evolution of network protocols and agent swarming

•  Wireless protocols for Intelligent Transportation System applications 

Service Layering for Enabling Collaborative SCAV Applications: The developed services for SCAV and their dependencies are depicted in the following diagram. The  SCAV mobile sensor system comprises the following four services: a) wheel or track based locomotion, b) multimodal onboard sensing, c) both ad hoc and access point oriented wireless networking, and d) infrastructure assisted indoor localization using Radio Frequency (RF) and Ultra Sound (US) signals. A point-to-point navigation service is developed using the localization and the SCAV locomotion services. A framework of Bayesian filtering has been incorporated to tackle with the intrinsic localization errors contributed by noises such as ambient RF and US interference, vehicle vibration, and inaccuracies introduced by the localization hardware components. Additional inaccuracies are caused when the localization intervals are too large to capture accurate locations of the moving sensors. A Position-Velocity-Acceleration (PVA) linear Kalman Filter has been used for smooth navigation in the presence of those localization errors. Finally, a set of collaborative sensor applications have been developed using the ad hoc networking and sensing services offered by the SCAV systems.

SCAV Self Localization: The components of the SCAV localization system are shown in the figure below. When a self-localization is needed, a SCAV computes its absolute distances from a number of static localization beacons pre-installed at known coordinates. Distance from a beacon is computed by measuring the time difference of arrival (TDOA) between an ultrasonic and a 433MHz RF signal simultaneously transmitted by the beacon. Once distances to a sufficiently large number of location beacons are computed, the SCAV computes its own coordinates using the distance values and the known coordinates of the relevant beacons.

Localization Performance: Performance of SCAV localization is presented in the following figure, which shows the absolute error represented as the difference between a SCAV’s actual coordinate and its location estimated by the localization mechanism described above. The primary source of this error is from faulty distance computation caused by the inaccuracies of the CRICKET hardware used as the beacons as well as listeners within the SCAV units. Other sources of errors were identified as various ambient RF and US interferences, and high frequency mechanical vibration contributed by the SCAV motors. The measurements were taken by statically positioning a SCAV at different locations in the 3mx3m test-bed.