This thesis reports the design and control of the ??GreenWeeder??, a non-herbicidal autonomous weeding robot, in order to autonomously track crop rows for weeding through electrocution in the inter-row space. The four wheel mobile robot platform was designed and built with a motorised Ackerman steering system allowing the robot to steer up to 30 degree left and right. It was also equipped with an electronically geared rear wheel drive, a pair of stereo cameras, a SICK LMS-291 laser range finder to localize itself with respect to the crop rows, a GPS system for obtaining the robot position in the field and a long-range communication system for tele-supervision by operators. The first prototype of the robot electrocution system was also designed and constructed to ignite 22kV electrical arcs to destroy weeds. Its operation was tested in the research field of the University of Sydney and the results of this experiment were analysed to improve the efficiency of this first prototype. An improved prototype of the electrocution system was then constructed and attached to a cradle extending out at the back of the mobile robot platform. The testing of this improved prototype was conducted at Lansdowne farm, a research field of the University of Sydney. After the construction of the robot platform, the robot control was considered with the demands for robot localization with respect to crop rows, an autonomously tracking control system and a manual control mode in order to take the robot to transportation vehicles. Firstly, the robot localization was accomplished by utilizing SICK LMS-291 laser range finder sensor for the sensing and perception of the robot. Secondly, the robot control system was developed with a PID controller, a second order model of the robot system and a first order filter. The PID controller is in the standard form with the filtered derivative and the PI part being in automatic reset configuration. The second order model was identified using Matlab System Identification toolbox based on the robot kinematic analysis. The first order filter is utilized for filtering out the lateral deviations of the robot with respect to the crop rows received from the SICK laser sensor. A Simulink simulation model of the robot control system was also built in order to fine-tune PID and filter parameters. Thirdly, the manual control mode of the robot was produced. In this mode, a joystick can be attached to a notebook to wirelessly drive the robot around or it can be plugged into a USB port at the back of the robot to drive it without the notebook. After the robot control was implemented and simulated, some experiments were conducted with the robot autonomously tracking a strip of reflective tape mimicking a crop row stuck into the ground of a laboratory. Depending on distances from the row assigned to the controller, the robot tried to keep those distances away from the row. The results showed the lateral errors of the robot with respect to the row were approximately 4.5 cm which were sufficient for our current agricultural application.

Peas dominate New Zealand grain legume production and they are a major export crop. However, weeds are a major problem particularly under organic production, where the use of synthetic chemicals is prohibited. To address this limitation, a research program to study weed control in peas was done to provide both conventional and organic farmers a sustainable weed management package. This was done through three field experiments over two growing seasons, 2006/07 and 2007/08. Experiment 1, (2006/07) evaluated the effect of 50, 100 and 400 plants m² on crop yield, and weed growth of Aragon, Midichi or Pro 7035 with and without cyanazine. Experiment two explored the physiology of two pea genotypes, the leafed (Pro 7035) and the semi leafless (Midichi) sown at three dates. A herbicide treatment was included as a control. In the third experiment Midichi, was used to investigate the effect of different pea and weed population combinations and their interaction on crop yield and weed growth. All crops were grown at Lincoln University on a Templeton silt loam soil. In Experiment one, herbicide had no effect on total dry matter (TDM) and seed yield (overall mean seed yield 673 g m²). There was also no significant difference in mean seed yield among the pea genotypes, Aragorn, Pro 7035 and Midichi, (overall mean, 674 g m²). The lowest average seed yield, 606 g m² was from 400 plants m² and the highest, 733 g m², from 50 plants m², a 21% yield increase. A significant herbicide by population interaction showed that herbicide had no effect on seed yields at 100 and 400 plants m². However, cyanazine treated plots at 50 plants m² gave 829 g m² of seed. This was 30% more than the 637 g m², from plots without herbicide. In Experiment 1 pea cultivar and herbicide had no significant effect on weed counts. In Experiment 2 the August sowing gave the highest seed yield at 572 g m². This was 62% more than the lowest yield, in October. Cyanazine treatment gave a mean seed yield of 508 g m², 19% more than from unsprayed plots. There was a significant (p < 0.05) sowing date x genotype interaction which showed that in the August sowing genotype had no effect on seed yield. However, in September the Pro 7035 seed yield at 559 g m² was 40% more that of Midichi and in October it gave 87% more. Weed spectrum varied over time. Prevalent weeds in spring were Stachys spp, Achillea millefolium L., and Spergular arvensis L. In summer they were Chenopodium album L., Rumex spp, Trifolium spp and Solanum nigrum L. Coronopus didymus L., Stellaria media and Lolium spp were present in relatively large numbers throughout the season. In Experiment 3 seed yield increased significantly (p < 0.001) with pea population. Two hundred plants m² gave the highest mean seed yield at 409 g m² and 50 plants m² gave the lowest (197 g m²). The no-sown-weed treatment gave the highest mean seed yield of 390 g m². This was due to less competition for solar radiation. There was no difference in seed yield between the normal rate sown weed and the 2 x normal sown weed treatments (mean 255 g m²). It can be concluded that fully leafed and semi-leafless peas can be sown at similar populations to achieve similar yields under weed free conditions. Increased pea sowing rate can increase yield particularly in weedy environments. Early sowing can also increase yield and possibly control problem weeds of peas (particularly Solanum spp), which are usually late season weeds. Herbicide can enhance pea yield but can be replaced by effective cultural methods such as early sowing, appropriate pea genotype and high sowing rates. Additional key words: Pisum sativum L., semi-leafless, fully leafed, cyanazine, pea population, weed population, sustainable, TDM, seed yield, weed, weed counts, sowing date, weed spectrum, seed rates.

Examensarbete. / Delrapport i FoU-projekt S 01-3284/08 "Skötsel av vegetation inom banområde"

Thesis--Virginia Polytechnic Institute and State University. / Also available via the Internet.

Research on the origins of agriculture in western Asia has placed great emphasis on the location and pace of domestication. However, much less attention has been given to reconstructing the specific nature and social implications of early cultivation practices across the agricultural transition, and to the potentially varied land management strategies involved. By employing a functional ecological approach to the interpretation of arable 'weed' taxa associated with early cultivars, this research addresses this gap in archaeobotancial research by enabling detailed analysis of the growing conditions and farming methods involved in early plant cultivation in western Asia. The core methodology analyses the functional ecological attributes (e.g. leaf area and thickness; canopy dimensions; stomatal density and distribution) of the relevant arable weed taxa isolated from archaeological contexts to determine the specific growing conditions of early crops and hence the nature of management practices. Functional attributes are morphological or behavioural characteristics that predict species' potential in relation to major environmental variables, such as soil productivity, disturbance and moisture. Statistical analysis incorporating these attributes is used to explore variation amongst early cultivation contexts and compare them with weed survey data from contrasting (semi-)arid modern regimes, including a recent study of traditional cereal farming in Morocco. Ecological 'signatures' were determined using the isolated weed dataset from four well documented and contextually rich Pre-Pottery and Pottery Neolithic sites, which were strategically selected to explore agricultural strategies from its initial stages through to the established Neolithic, as well as to exploit detailed sample-by-sample data and extensive in situ deposits. The sites investigated are PPNA Jerf el-Ahmar and PPNA/EPPNB Dja'de in northern Syria, PPNB Tell Aswad in southern Syria, and PPN-PN Çatalhöyük in Central Anatolia. Refined identification of selected weed genera at these sites enabled more accurate indications of their arable ecologies. The results generated suggest that early farming practices were highly variable within sites, reflecting the specific affordances of local climate and surrounding landscapes, but relatively labour-intensive in comparison with later urban agrosystems. Furthermore, there are indications for greater cultivation intensity over time, as households became more autonomous.

Master of Science / Department of Agronomy / J. A. Dille / Weed control in grain sorghum has always presented a challenge to producers in the semi-arid Great Plains. Cultural control tactics such as narrowing of row spacings and increasing seeding rates can be effective control methods. The objective of this research was to determine the row spacing and seeding rates that maximizes yield while suppressing weeds. Grain sorghum row spacings of 25, 51, and 76-cm and seeding rates of 75,000, 100,000, 125,000, and 150,000 seeds ha[superscript]-1 were evaluated in Kansas at Beloit and Manhattan in 2013 and Beloit, Manhattan, and Hays in 2014. Grain sorghum growth and yield response were measured in response to natural weed communities. After evaluation, Beloit was considered a low weed pressure site while Manhattan and Hays were considered to be moderate and high weed pressure sites, respectively. Grain sorghum biomass was different while weed biomass was consistent across row spacings. Yield loss equations and profit functions were derived to determine the amount of grain yield and $ ha[superscript]-1 loss from each of the three locations. Yield and profit lost was greatest amongst weedy observations. Results indicated that grain sorghum grown on wide row spacings and seeding rates of 125,000 seeds ha[superscript]-1 out yielded all other treatments under a low weed pressure site (Beloit) and narrow row spacings out yielded wider spacings in moderate and high weed pressure sites (Manhattan and Hays). These results imply that a Kansas grain sorghum producer should evaluate potential weed pressure before determining a final row spacing and seeding rate.

Made available in DSpace on 2014-06-11T19:28:31Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-07-03Bitstream added on 2014-06-13T20:58:20Z : No. of bitstreams: 1 yamauti_ms_me_jabo.pdf: 351648 bytes, checksum: 2354f2240d26120716bf47e94b215e44 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O presente trabalho foi conduzido com o objetivo de se determinar o efeito da adubação sobre o período anterior à interferência (PAI) das plantas daninhas na cultura do amendoim, cv. Runner IAC 886, semeada no sistema convencional. Os períodos de convivência estudados foram: 0, 7, 15, 20, 30, 45, 55, e 126 dias após a emergência (DAE), totalizando oito tratamentos, que foram dispostos em blocos casualizados, com quatro repetições, em dois níveis de adubação (0 e 150 kg ha-1 da formulação 00-20-20 (N-P-K)) caracterizando-se como dois experimentos. A comunidade infestante foi composta predominantemente por Cyperus rotundus, Eleusine indica, Indigofera hirsuta, Portulaca oleracea e Sida rhombifolia nos dois níveis de adubação. A cultura pode conviver com a comunidade infestante até 17 e 15 DAE sem sofrer redução significativa na produtividade (PAI) para 0 e 150 kg ha-1, respectivamente. A interferência das plantas daninhas reduziu, em média, 30% a produção do amendoim. / The experiment was carried out to determine the effect of fertilization on the period before weed interference (PBI) on weeds in peanut Runner IAC 886 in a conventional till system. The periods of weed control studied were: 0, 7, 15, 20, 30, 45, 55, and 126 days after emergence (DAE), totaling eight treatments, arranged in a randomized block design, with four replications and two levels of fertilization (0 and 150 kg ha-1 of formulation 00-20-20 (N-P-K)) characterizing as two experiments. The weed community comprised species with Cyperus rotundus, Eleusine indica, Indigofera hirsuta, Portulaca oleracea e Sida rhombifolia being the most prominent in all levels of fertilization. Peanut grew along with the weed community up to 17 and 15 days after emergence, without significant yield loss (PBI) to 0 and 150 kg ha-1, respectively. Weed interference during the full crop cycle reduced, in average, 30% of peanut productivity.

Orientador: Geraldo Nunes Silva. / Banca: Weldon A. Lodwick. / Banca: Valeriano Antunes de Oliveira. / Resumo: O propósito desta pesquisa é o estudo de dois modelos de otimização dinâmica para a aplicação seletiva de herbicida , sendo um modelo de otimização multi-objetivo , no qual maximiza o lucro e minimiza a resistência , e o utro mono-objetivo da otimização da concentração de mistura de herbicidas, no sistema anual de colheita da cultura do milho para o período de 5 e 10 anos. A densidade de sementes no solo no início do plantio e a frequência de alelos são tomados como variáveis de estado . A variável de controle é expressa na função de dose-resposta. Os modelos de otimização levam em consideração a diminuição da eficiência do herbicida ao longo do tempo , causada pela evolução da re-sistência da planta daninha . O objetivo é maximizar o lucro num período pré-determinado e minimizar a evolução da resistência. O problema de otimização dinâmica multi-objetivo foi resolvido via abordagem−restrito. O problema resultante e o problema de otimização da concentração de misturas de herbicidas foram resolvidos por programação não-linear via o método ASA_CG. Para os dois casos estudados, os resultados das simulações numéricas a presentam uma estratégia ótima da aplicação de herbicidas, para... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The purpose of this research is the study of two dynamic optimization models for the selective application/mixture of herbicide in the annual system of corn crop harvest for period of 5 and 10 years. The density o f seeds at the beginning of the planting and allele frequency are taken as state variables. The control variable is given by the dose-response function. The optimization models take into account the decreased efficiency of the herbicide overtime caused by the evolution of weed resistance. The goal is to maximize profit in a predetermined period and minimize the resistance evolution. The dynamic optimization problem of multi-objective approach was solved by the e-constraint method. The resulting problem and the problem of dynamic optimization of the concentration of the herbicide mixture were solved by non- linear programming with the conjugate gradient method combined with the method of projected gradient. Results of numerical simulations provide an optimal strategy of selective application of herbicides to control weed infestation by the Bidens subalternans. Two models studied, the first considers only... (Complete abstract click electronic access below) / Mestre

Master of Science / Department of Agronomy / Johanna A. Dille / An opportunity for postemergence (POST) grass weed control has recently been approved with ALS-resistant grain sorghum, however, grass weed emergence timing and crop tolerance to grass competition are not well understood. To address the importance of POST application timing, a critical period of weed control (CPWC) for grass competition in grain sorghum was developed. Field experiments were established near Manhattan and Hays, KS in 2016 and 2017, and near Hutchinson, KS in 2017 to determine the CPWC. Each site provided a different grass species community. A total of ten treatments were included, with four treatments maintained weed-free until 2, 3, 5, or 7 weeks after crop emergence, four treatments receiving no weed control until 2, 3, 5, or 7 weeks after crop emergence, and two treatments were maintained weed-free or weedy all season. Treatments did not influence grain yield at Hutchinson because of a lack of season-long weed emergence. At Hays the CPWC began at crop emergence and ended 28 days later. At Manhattan the CPWC began 27 days after emergence and continued through grain harvest. The CPWC in grain sorghum depends on rainfall and competitive ability of the weed species. The start of the CPWC began when weeds emerged, thus a POST application should be targeted 14 to 21 days after emergence of grain sorghum. Emergence and development of large crabgrass, barnyardgrass, shattercane, and giant, green, and yellow foxtails were studied near Manhattan, KS after seeding on April 11, 2017. Barnyardgrass had the longest duration of emergence, beginning at 180 GDD after seeding and continuing through July. Large crabgrass had the shortest duration of emergence from 325 to 630 GDD after seeding. In general, all grasses began to emerge in late April and most species completed 90% emergence by early June. Grain sorghum is typically planted at this time, so grass weed control prior to planting is critical. Palmer amaranth is a troublesome weed in double-crop grain sorghum production fields in Kansas. The presence of herbicide-resistant populations limits options for weed management. Field experiments were conducted to evaluate 14 different herbicide programs for the management of atrazine-resistant Palmer amaranth in double-crop grain sorghum at Manhattan and Hutchinson, KS in 2016 and 2017. Programs included eight PRE only and six PRE followed by POST treatments. Programs that had very long chain fatty acid-inhibiting herbicides provided greater control of atrazine-resistant Palmer amaranth by three weeks after planting sorghum. Programs of PRE followed by POST provided greater control of both atrazine-resistant and -susceptible Palmer amaranth by eight WAP compared to PRE alone. These results illustrate the value of residual herbicides, as well as an effective postemergence application, in double-crop grain sorghum. Early season grass and Palmer amaranth control with the use of residual herbicides such as very long chain fatty acid-inhibitors provide a competitive advantage to grain sorghum. Utilizing weed emergence patterns to time effective POST applications, in unison with residual herbicides, will provide season-long weed control in Kansas grain sorghum fields.

Maize is the third most important cereal crop in the world and many subsistence farmers are practicing intercropping of it with legumes due to land scarcity and in order to enhance production. Intercropping system is being practiced in may areas of South Africa mainly in the Limpopo province. The objective of the study was to evaluate the crop responses to intercropping maize with different growth length cowpea cultivars. The effects of intercropping on weed growth, maize and cowpea growth and yield components were investigated. The experiment was conducted during the 2005/06 and 2006/07 seasons at Bethlehem and Potchefstroom. Treatments were: maize sole, maize+PAN311 (short duration cowpea cultivar), maize+Glenda (medium duration cowpea cultivar), maize+Agrinawa (long duration cowpea cultivar) and sole plots of all cowpea varieties. Each plot was divided into two weed levels where all the plots were kept weed free for one month after planting, after which one half was left weedy and the other half weed free. Weed sampling was done within each weed treatment. Intercropping reduced maize LAI and plant height while time to physiological maturity was also reduced by weed infestation, especially under drier and warmer environments. Glenda and Agrinawa intercrops produced more nodules per plant under cooler and wetter conditions. Agrinawa produced the highest leaf and total DMY under sole crop conditions and this was significantly reduced by weed infestation. Different growth duration cultivars did not differ in their N2-binding abilities. Maize intercropping, especially with Glenda and Agrinawa, significantly reduced weed biomass. Maize sole crop under zero weeds had high grain yield compared to intercropping. PAN311 and Glenda sole crops under zero weeds produced higher yields under dry and warmer conditions, and cooler and wetter conditions, respectively, compared to intercropping. High cowpea grain yields were strongly correlated to more seeds per pod and larger pod lengths and number of pods per plant especially for Glenda. No intercropping advantage compared to sole cropping was observed (total LER < 1). This implies that maize and cowpeas must rather be planted as sole crops for better yields under wetter and cooler, and warmer and drier conditions. Copyright / Dissertation (MSc(Agric))--University of Pretoria, 2009. / Plant Production and Soil Science / unrestricted