25813 research outputs found
Sort by
Monty: design of a wireless, battery-powered snakelike robot
No full textThe pursuit of a mobile robot capable of effectively navigating cluttered, unstructured environments is a goal many researchers seek to achieve. Whether for unmanned exploration, or to aid in safe search and rescue efforts, the applications for effective unstructured navigation are numerous. Conventional robots achieve locomotion largely through the use of motorised wheels or tracks, which present a simple solution. However, where these conventional robots fail is in navigating difficult, rugged terrain. Therefore, a new approach was taken, in that researchers look to nature for inspiration. Legged movement presents an interesting and highly dextrous method, but when applied to robotics, the kinematics involved can quickly become excessively complicated. Another alternative is to mimic the versatility and hyper-redundancy of biological snakes, i.e. snake-like robotics. Snake-like locomotion is characterised by undulation by means of a long, slender body, relying on friction and exploiting the terrain profile to generate momentum. A robot capable of effectively harnessing this dextrous motion could prove invaluable.
This thesis will present Monty, a wireless, low-cost snake-like robot constructed of eleven reconfigurable chained modules equipped with frictional scale panels, and a head module with a front mounted camera. The robot is capable of 3D locomotion, in particular, sidewinding and lateral rolling, and basic obstacle detection. The design was adapted from Noodle, where limitations were found and improved, and additions were made to facilitate the completion of new objectives. In particular, the objectives for this research were to create a wireless design, powered by batteries, and capable of 3D locomotion gaits and obstacle detection via a camera.
Each module body is constructed of 3D printed resin for a durable and inexpensive approach, with the overall price for each module coming to 181 NZD. This cost is largely due to the servo motors and electronics chosen for use in each module. To provide actuation, the existing series elastic actuator (SEA) assembly was retained from Noodle, consisting of a servo, polyurethane elastic element acting as a torsional spring, and an angular encoder to measure deflection in the joint. The use of the SEA has several benefits; in particular, this allows for compliance, energy saving and torque sensing, which can be used to enable adaptive control feedback. The modules can also be equipped with three different ventral scale panels, which provide desirable friction qualities to complement each gait. Power is provided independently to each moving, by means of a 450 [mAh] battery, which supplies power for approximately 43 [min].
In addition to the eleven identical body modules, Monty is equipped with a dedicated head module to house a Raspberry Pi, which provides wireless communication and image processing from the camera. Wireless communication to the users personal computer (PC) is achieved with the secure shell (SSH) protocol, through the local WiFi network. The Raspberry Pi reads the users input and generates the gait wave pattern to be sent to the first module using the serial universal asynchronous receiver transmitter (UART) protocol, which is then relayed to all other modules with the controller area network (CAN) protocol. The Raspberry Pi is also used to process the video feed from the front mounted camera. A basic computer vision program is used to identify possible obstacles in the image, which it does with an average accuracy of 87%.
To facilitate 3D locomotion, the modules were arranged in a pitch-yaw configuration, in which each alternate module is rotated 90°. Through testing, the robot was found to be capable of effective movement with the sidewinding gait, at an average speed of 0.125 [m/s], and lateral rolling, at a average of 0.04 [m/s].
Validated through testing, Monty successfully achieved all the objectives outlined in this thesis, as a modular, wireless, snake-like robot with 3D locomotion and obstacle detection capabilities. Although it has some limitations, Monty represents a practical advancement toward creating a versatile robot capable of navigating challenging environments
Tree Canopy Cover in Whakatāne 2023
No full textThis report was prepared by Professor Justin Morgenroth and Dr. Ning Ye at the School of
Forestry, University of Canterbury. The aim of this report is to provide local authorities in New
Zealand with a basic understanding of the urban tree canopy cover within their cities and towns
Non-invasive optical methods for the detection of blood analytes in the near infrared spectrum.
No full textSeveral diseases and areas of medicine would benefit from non-invasive blood analyte
monitoring. Specifically, blood glucose (BG), blood creatinine (BC), and blood lactic acid
(BL) for management of diabetes mellitus, assessment of kidney function for chronic kidney
disease, and diagnosing and monitoring septic shock. Currently, all methods for measurement
of BG, BC, and BL are invasive.
As of 2023, 323,700+ people in New Zealand are living with diabetes mellitus, a disease
causing a relative deficiency of insulin, the hormone that regulates BG concentration.
Semi-invasive continuous glucose monitors (CGMs) provide continuous monitoring for up to
two weeks, but are not funded for up to 90% of diabetics in New Zealand. CGMs can be
privately purchased but are prohibitively expensive, restricting access and leading to low
compliance with BG monitoring. Low compliance with BG monitoring correlates directly
with an increase in poor health outcomes for patients living with diabetes.
Kidney disease can arise as a complication of poor management of diabetes, and in
other cases such as kidney failure in the intensive care unit. Monitoring the concentration
of creatinine in blood (BC) and urine allows calculation of estimated globular filtration rate
(eGFR) to monitor kidney function. Currently, eGFR is only monitored in a hospital setting
using lab-based assays, and no non-invasive methods are available.
Blood lactic acid concentration is a key blood analyte which can be indicative of septic
shock, leading to coma and death if untreated. BL monitoring in clinical settings is done
discretely via lab-based assay. This increases the potential for septic shock to progress in
patients between readings, increasing morbidity and mortality. In addition, a rise in BL can
be induced through intense exercise. Lactate threshold testing monitors this BL rise and is a
common method for determining training zones and improving athletic performance. There
are currently no non-invasive methods to measure BL in medical or athletic settings.
All current methods for measurement of BG, BC, and BL are invasive and carry a risk
of infection, are painful and inconvenient, and are expensive. Point of care devices can
cost up to NZD450 for a 50 pack of the consumable test strips. For these
reasons, patient compliance to regular testing is low. The aim of this thesis is to develop
a platform technology for non-invasive blood analyte detection using commercially available
components. The Sweet As sensor, version 2 (SAv2), is designed as this platform technology,
detecting glucose, creatinine, and lactic acid using discrete near-infrared (NIR) spectrometry.
The spectral response curves for for seven NIR light emitting diode (LED) wavelengths
used in this thesis (1050 - 1650 nm) and a NIR photodiode (PD) (900 - 1700 nm) were developed for analysis of LED-LED analyte sensing. The LEDs could be used on a noninvasive blood analyte sensor in place of a PD. However, sensitivity and cost of the NIR-LEDs
does not provide a sensing advantage over the PD, which has a spectral response functionally
similar to the 1650 nm LED with enhanced detection of the lower wavelengths. For these
reasons, the PD with absorption range 900 - 1700 nm is used on the SAv2.
The absorption spectra for six compounds, glucose, lactic acid, creatinine, water, acetone,
and ethanol were developed using three lab-based spectrometers with difference measurement
methods. The absorption spectra developed do not align with the the literature. However,
the variation shown in the literature suggests absorption values may not be comparable
across different spectrometers. Wavelength combinations for the SAv2 were selected for
the detection of each blood analyte using the absorption spectra developed. Combinations
1450/1460, 1550, and 1650 nm, or 1200, 1450/1460, and 1550 nm are suggested for detection
of glucose. Combination 1050, 1200, and 1300 nm are suggested for detection of lactic acid,
and 1200, 1450/1460, and 1550 nm wavelengths for detection of creatinine.
The volume and molar percentage of molecules in 1 litre of blood was estimated using
physiological concentrations. Specifically, whole blood was approximated to contain 50%
water, 40% haemoglobin, and 10% other components by volume. At concentrations of 5.0
mmol/L for glucose, 2.0 mmol/L for lactic acid, and 0.8 mmol/L for creatinine, these analytes
comprise less than 0.2% of blood volume and under 0.02% of total molecules. Despite sharing
the same molar percentage as creatinine (0.0028%), haemoglobin constitutes 40% of blood
volume due to its large molar mass, suggesting that molecule size contributes to its optical
detectability in pulse-oximitry. These findings imply that the small molecular size of glucose,
lactic acid, and creatinine significantly limits detection via non-invasive optical methods,
despite their molecular abundance relative to haemoglobin.
A series of 3D Monte Carlo (MC) simulations investigated the effects of varying blood
analyte concentrations on the light-tissue interaction of NIR wavelengths between 1050 -
1650 nm. The simulations used an 8-layer tissue model and the three blood analytes at three
concentrations. The mean photon depth and pathlength ranged from 1.1 - 1.3 mm and 7.3
mm - 8.9 mm respectively, with small deviations observed across the simulations. There was
minimal deviation in normalised absorption and normalised fluence rate with no correlation
to analyte concentration, suggesting changes in analyte concentration have a minimal effect
on NIR absorption in the tissue. The pathlengths calculated from these MC simulations
were used in later experiments and clinical testing using the SAv2 on humans.
The transmission of six NIR wavelengths through varying thicknesses of porcine tissue
showed light transmission decreases with increasing wavelength. The experiments confirmed
the NIR wavelengths can reach the depths and pathlengths calculated in the MC simulations,
and can likely reach a tissue depth of up to 3.3 mm with a pathlength of up to 28.0 mm in
a reflectance configuration with 2.0 mm spacing. The experiments also showed the digital
timing method for data acquisition improved detection at lower incident light levels.
Varying analyte concentrations in aqueous, and acetone solutions, and whole human
blood were analysed using the SAv2 sensor. The sensor response was generally more pronounced
in blood than in water, particularly at lower wavelengths, suggesting differences
in light interactions with the blood components and water play a crucial role in sensor response.
Creatinine showed the largest percent change in sensor response across a wide range
of concentrations. The change in sensor response to a change in BG and BL concentration
was smaller and variable.
Three studies assessed the ability of the SAv2 to detect BG and BL concentrations separately
on humans. Trial 1 presents the SAv2.1 for detection of BG with emitting wavelengths
of 1450, 1550, and 1650 nm, with 1650 nm LEDs as detectors. The sensor showed a weak
correlation between estimated and reference BG levels, yielding an R2 value of 0.038. Several
limitations of the study were identified, most notably poor signal quality, which resulted in
a data loss of 41.0%.
Trial 2 evaluated the SAv2.3.1 and SAv2.4.1 devices, using wavelength combinations of
1050, 1200, and 1300 nm, and 1200, 1460, and 1550 nm, respectively, with PDs for detection.
The use of PDs significantly enhanced signal quality and no data was discarded. Applying
individual participant calibration improved the correlation accuracy for both devices giving
an R2 value of 0.539 and 0.285 for the SAv2.3.1 and SAv2.4.1, larger than the pilot study.
Despite these improvements, the overall correlation between estimated and reference BG
values remained relatively low.
In Trial 3, the SAv2.4.2 was tested for its ability to detect changes in BL concentration,
using emitting wavelengths of 1050, 1200, and 1300 nm with PDs for detection and a covering
over the sensing array for protection from sweat. Due to poor signal quality, 45.0% of the
data was excluded from analysis. After individual calibration to each participant, a moderate
correlation was observed with R2 = 0.601, but the mean absolute relative difference (MARD)
between the estimated BL and paired reference measurement was high at 52.41%. These
results suggest the SAv2 is not capable of reliably detecting changes in BL concentration.
Potential criteria for an analyte to be detected were drawn from the findings of this
research and include: 1) the analyte should not contain an -OH bond; 2) the analyte should
have distinct features on the absorption spectra that distinguish it from the absorption
spectrum of blood; 3) the percent volume of the analyte in the blood should be high enough
such these distinct features can be detected at physiologically relevant analyte concentrations;
4) wavelengths outside the NIR spectrum should be utilised to avoid high absorption from
water; and 5) a transmission sensor may provide advantages over a reflectance sensor if used
in the correct location on the body, ensuring the light has passed through pulsatile tissue.
Overall, this thesis provides novel contributions to the understanding of the spectral
response of NIR LEDs, the absorption spectra of glucose, creatinine, and lactic acid, and
non-invasive NIR sensing methods for these three blood analytes. Results show the SAv2 is
not a feasible avenue for analyte detection in the blood at this stage
Modelling seawater intrusion in New Zealand’s alluvial aquifers that extend offshore.
No full textCoastal aquifers provide fresh water for over one billion people around the world.
This resource is under increasing pressure from the effects of climate change and
population growth. Consequently, the risk of seawater intrusion, which results in
the salinization of fresh groundwater, is heightened. In New Zealand, coastal alluvial
aquifer systems form many of our most important water resources. In these systems,
it is common that the interface between fresh and saline groundwater is located offshore.
Regardless, they are still susceptible to seawater intrusion. This thesis aims
to improve the understanding of seawater intrusion in New Zealand’s alluvial aquifer
systems that extend offshore. This has been achieved through variable-density numerical
groundwater modelling. Three objectives have been addressed. The first was
to conceptualize seawater intrusion in a simplified confined aquifer that extends offshore
and quantify the effects of key hydraulic properties. The second was to assess
whether preferential seawater intrusion caused by heterogeneity can occur in alluvial
aquifers, based on a case study of the Central Canterbury Plains. The final objective
was to investigate the influence of seafloor pockmarks on submarine groundwater
discharge and seawater intrusion, based on the Lower Hutt Valley aquifer system.
Results showed that seawater intrusion could occur vertically or horizontally into
the confined aquifer that extended offshore. Aquifer hydraulic conductivity increased
the amount of salinization, and aquitard vertical hydraulic conductivity had a mixed
effect. Preferential seawater intrusion was found to occur in alluvial aquifers, resulting
in variable salinization rates and significant alongshore variation in the amount
of seawater intrusion. A comparison of heterogeneous aquifer models and equivalent
homogeneous models based on the effective hydraulic conductivity, including dispersion,
found that homogeneous models underestimated the velocity of the interface
by up to ten times. Pockmarks were found to enhance submarine groundwater discharge,
with discharge rates being greatest when conduits were included in models.
Models with conduits also experienced the greatest amount of seawater intrusion
under aquifer depletion. Comparing three different geological models of pockmarks
to observations in the study site, all of the models could explain some but not all of
the measurements. Groundwater researchers and practitioners should consider the
possible effects of both vertical and horizontal preferential seawater intrusion when
studying alluvial aquifer systems that extend offshore
Bridging the adoption gap in primary sectors : a systemic technology adoption model.
No full textInnovation adoption in primary sectors—agriculture, horticulture, forestry, and aquaculture—is essential for addressing pressing global challenges, including climate change, resource degradation, and food security. However, a persistent gap exists between innovation potential and actual implementation, with many promising technologies failing to achieve widespread adoption despite substantial research investments. This paper presents the Systemic Technology Adoption Model (STAM), a conceptual model that addresses critical gaps in adoption theory by integrating four quadrants: technologies, users, finance, and institutions. STAM explicitly recognizes adoption as a systemic process requiring alignment across multiple dimensions. The model’s distinctive contribution lies in its emphasis on inter-quadrant relationships, revealing how variables across different domains interact, compound, and cascade to create either enabling conditions or barriers. Through a test case, we illustrate how the model can enable practitioners to proactively identify potential adoption barriers early in the innovation development process, revealing when barriers in multiple quadrants compound to create obstacles, when cascade effects amplify constraints across the system, and where strategic interventions can address multiple barriers simultaneously. We discuss theoretical contributions and practical implications for practitioners and policy designers, highlighting how STAM could provide stakeholders with a tool for designing more effective adoption strategies
A highly-tunable novel Rocking Tube Damper (RTD).
No full textSelf-centering rocking systems with passive energy dissipation devices have been studied extensively in the literature because of their low-damage seismic behaviour and resilience. The rocking mechanism is typically a feature of the structural system, and some form of supplemental energy dissipation system is often added to the structure to absorb energy and limit displacement response.
This paper proposes a new Rocking Tube Damper (RTD) which incorporates a gap opening rocking mechanism into a scalable and non-invasive supplemental damping device. The self-centering rocker is preloaded with a stack of partially compressed conical spring washers. These conical spring washers provide elastic deformation capacity, and together with the applied preload produce a largely bi-linear elastic response that creates an initial resistance to motion and controlled inelastic behaviour.
The RTD has been previously modeled to establish its force-displacement response, showing an ideal bilinear elastic hysteresis with zero residual deformation. A prototype of the self-centering rocker has been fabricated and experimentally tested under cyclic loading. The experimental response matches the broad behaviour expected from the numerical model. However, initial experimental test results exhibited some inherent energy dissipation from friction than was not included within the mathematical model. Suggestions are made for ongoing work to further investigate the full range of response behaviors of the RTD concept
Evaluating the implementation of youth audit recommendations in Christchurch: an assessment of stakeholder engagement with ReVision’s youth-relevant design audits
No full textThis community-partnered thesis explores the impact of ReVision Youth Audits in promoting youth-friendly community spaces in Christchurch, a city undergoing long-term urban transformation following the 2010–2011 earthquakes. In partnership with ReVision, a not-for-profit organisation facilitating youth-led audits of public and community spaces, this research examines how audit recommendations have been implemented by organisations responsible for 23 previously audited sites. Using a mixed-methods approach, including an online stakeholder survey (n = 16) and semi-structured interviews (n = 2), the study identified variation in implementation outcomes, with non-profit organisations reporting higher adoption levels than local government entities. Stakeholders reported that commonly implemented recommendations included enhanced lighting, inclusive signage, additional seating, and youth-focused amenities such as murals, free Wi-Fi, and gender-neutral toilets. The average youth-friendliness score increased from 4.7 to 7.5 out of 10 following implementations, reflecting tangible improvements in accessibility, inclusivity, and youth engagement.
Despite these gains, several barriers limited full implementation. Local government stakeholders cited procedural delays, regulatory frameworks, and funding cycles tied to long- term planning. At the same time, non-profits stakeholders faced constraints such as property ownership and limited influence over shared spaces. Challenges related to timing, staffing capacity, and the absence of follow-up mechanisms were also reported. Stakeholders recomended integrating youth input in the design process earlier, as several audits occurred after key planning phases. Feedback on the audit process was largely positive, with high ratings for the clarity of recommendations and the tool's credibility. However, stakeholders advocated for refinements when recording the audit recommendations to capture young people's lived experiences better and sustain youth involvement beyond the initial audit phase. The research demonstrates that the ReVision Youth Audit framework contributes to meaningful improvements in public spaces especially for youth and reinforces the value of youth-informed urban design. This research provides practical guidance for enhancing youth engagement in urban planning and improving the long-term utility of participatory audit frameworks, based on an analysis of both the factors that enabled and those that constrained the implementation of audit recommendations
